In June I wrote a short piece about the difficulty I was under trying to understand why the equation
GDP = C + I + G + (X-M)
is a simple accounting identity, a self-evident truth rather than an assumption made by economists. Were it wrong, we’d have to dismiss 500 years of double-entry bookkeeping, which is obviously inconceivable.
To recap: the equation states that a country’s GDP equals the total of private consumer spending (C), private investment (I), government spending (G), and the trade balance of exports less imports (X-M). From this economists derive the result that if GDP is to be kept constant, then reducing public debt implies either increased personal borrowing or a higher trade surplus or both. Some countries – Germany is the prime example – want to reduce the deficit by exporting. In the UK, however, since the trade balance is not likely to be turned around quickly, the government has to hope for a new private debt boom. Current policy, therefore, involves cajoling the banks into lending as much to the private sector as the government wishes to cut from public indebtedness.
We’ll see what happens, but in the meantime, I want to come back to this nagging and somewhat nerdy question of accounting identities, because if this and similar equations are not unquestionable truths about the world, then a whole new range of policy options opens up. TINA might become TIARA.
In my previous post, I said:
“[Bookkeeping] is a closed system. Its truths are true by definition. Bookkeeping defines both the objects measured and how they are measured. In particular, it measures money or money equivalents. The relationship of these measured things with the real world is genuinely problematic.”
I’d like to develop that idea a little. At a practical level, I think the story is about completeness.
Future pension liabilities (both the basic state pension payable to all and the full pensions of public sector employees), contingent liabilities, payments under PFI contracts, net debts of publicly owned corporations, all manner of future public spending is not recorded as public debt. In the UK, pension liabilities alone have been estimated at 230% of GDP, and the figure for Germany is even higher.
The same applies to the private sector. Rental payments on long-term leases are expensed as paid. All sorts of participants in the economy hold derivative contracts, from small, business-related hedging to large or speculative punts. How these can be valued and reported is uncertain. Companies use off-balance-sheet vehicles. Rich individuals place assets offshore and take out a mortgage in the UK. We pay cash to the home help or, if we are lucky, offer room and board to the au pair. Charities receive commitments of future time from volunteers. Entrepreneurs invest their time developing a business idea. We could produce examples indefinitely.
In the corporate sector, accounting standard setters continually debate just how leases, long-term obligations or revenues, securities, stocks, sources of finance, currencies, and all the rest should be reported in company accounts. It’s notable how little comparable debate there is over how the public sector should report these things, as if the public accounts were either too big and complicated (or maybe not important enough) for ordinary citizens to bother with.
This measurement problem has two aspects. First, vast parts of the real economy are just not included in its published accounts. Sometimes parties wish to hide what they do or manage tax liabilities, but often this is just in the nature of things. No politically acceptable system could be designed that would report all private and public transactions with an impact on the national economy.
Secondly, even where transactions are there to be measured, we face uncertainty. For example, accounting for leases or derivatives requires intelligent and honest judgments of what is happening. The search is for a meaningful answer, not an eternally right one that can be added to all the other right answers. Accountants provide a salutary lesson. The profession creates more and longer standards, with international harmonisation where possible, but there is little public confidence that financial statements are fundamentally – taken as whole – more useful or reliable or incorruptible than 20 years ago. If anything, recent banking and other corporate collapses suggest the opposite is the case.
In a word, I doubt we can credibly measure any of the values in the ‘identity’. It leaks everywhere. Even if there is an identity, we can’t quantify its components. So why assert the identity when none of its terms is well defined, especially since the effects we expect from pulling different policy levers are just as uncertain?
There’s another, more epistemological level too. The argument based on the antiquity of double-entry bookkeeping is pleasantly rhetorical, but does it have merit? What theory of value is in play?
The magic of double-entry bookkeeping is that for every recorded transaction, I make at least two mutually cancelling entries in the ledger. If I pay out £1, I note that on the right-hand page of my cash book, and must then record £1 on the left-hand page of another ledger to say what it was spent on. The totals of the right-hand and left-hand columns are always, by definition, equal.
Those ledgers that describe unsatisfied claims I have or unsatisfied claims against me are summarised in the balance sheet; the profit and loss account summarises all the others. Accounting standards are entirely to do with deciding which document my ledgers are reported in and how.
Is this a helpful description of how value is made or lost in economic activity? Let’s try a thought experiment. Suppose I pay out £100 for goods or a service. It turns out that I am so happy with the outcome that I think it was worth £400. I also pay £100 for something else that doesn’t work out so well; in fact, I think it was a total waste. What I now want to do is record my expenses as I valued them: £400 + £0 = £400. For these, I gave out £200. In this model, success means being an expense maximiser, since that means I’ve spent wisely. But double-entry bookkeeping tells me my expenses must be precisely £200. How well I did goes by the board. A purely pecuniary model drives me instead to minimise expenses, since value for money is irrelevant to it.
More paradoxically, on the other side I should minimise my revenue, since that would be a sign that I’ve sold goods or services at a higher price than I valued them. The concept of profit or loss, certainly that of the profit maximiser, starts looking rather shaky. (This scenario has nothing, of course, to do with reducing tax liabilities, since in the experiment, tax would be assessed completely differently.)
But this isn’t just an intellectual game. What about mark-to-market valuation of securities; what about companies that buy a property and revalue it? What about those that buy an expensive corporate toy and then junk it because it’s useless? What about goodwill – the difference between what one company pays (itself ambiguous if securities are exchanged) to acquire another and the target’s apparent “value”, a difference that ends up being expensed over time, so increasing costs by the value of the thing acquired? What about brand value? This all looks very like what I’ve simplistically sketched out, only with a whole load of tap-dancing required to keep the double-entry in balance.
Wasted effort is one thing, but there may also be perverse consequences. Every time I change my view of the value of something, double-entry requires me to write the financial effect of that into at least two ledgers, whether or not anything real has happened. Hey presto! My revaluation suddenly generates a profit or a loss. Even allowing for the many complex rules around realised and unrealised, distributable and undistributable profits, we can see the potential for abuse, whether over bonuses, tax, or earnings. It rather looks as if double-entry bookkeeping, originally conceived as a system of error detection, might bizarrely be forcing those who rightly and justifiably want to reflect the real value to them of what they have bought and sold to construct all kinds of unreal values too, merely to keep the ledgers in balance. Are national accounts any less subject to abuse, errors, or structurally imposed fictions?
Perhaps the largest group of public obligations whose accounting and the politics behind it need to be brought out into the light and thoroughly overhauled are pension liabilities and payments, which I mentioned earlier. The estimation of the cash value of pension liabilities is technical and requires some major assumptions. Beyond that, how these liabilities are actually valued and reported, and how that information is used in economic management, is highly political. I hope to return to pensions another time.
Coming back to GDP = C + I + G + (X-M), it seems that the quantities included there are all subject to what auditors refer to as “fundamental uncertainty.” Furthermore, 500 years of double-entry bookkeeping may have led us to invent identities that mislead rather than enlighten, and to mistake an accounting definition for a transmission mechanism. We should focus on real value – however politically and subjectively difficult that is to ascertain – rather than on monetary cost and accounting purity. Then we might conclude that public debt reduction is possible without more net exports or personal borrowing, perhaps by redirecting expenditure to areas of greatest public value, since our evaluation of government spending varies over time and according to the things bought.
What on earth was the monk Luca Pacioli thinking as he wrote his mathematics textbook for the sons of fifteenth century Venetian merchants?