What is the story in the larger part of the world where productivity has risen but not so much as to allow the types of gains that have created the wealth of the developed world? In this chapter we are going to focus on three cases: South Korea, China and India. They each started late and had different degrees of success in raising their productivity; so much so that they are sometimes seen as examples of miracle productivity growth. But first we need to put their stories into the wider perspective of the global productivity pattern.

Most of the high-productivity economies today are members of the OECD. It was formed in 1961 as a “Western” group in the midst of the Cold War. Today it has 36 members with an average per capita nominal income of some $41,000. Unfortunately, the population of the OECD countries today makes up only 17 per cent of the world's population. High productivity is not an explicit membership requirement of the OECD. Politics matters too. This explains why Russia is still not a member. But the high productivity of the OECD economies makes them a huge force in the global economy. In Asia only Japan and South Korea are members. In Central and South America only Chile and Mexico. No African country is a member of the OECD.

Table 6.1 shows the productivity gaps which exist today. We are again measuring these in terms of GDP per person employed because the poorer economies lack good enough data to calculate GDP per hour worked. Some regional categories are ragbag ones. There is a big gap, for example, between the economies with oil in the Middle East and North Africa and those without it. Still, the contrasts are sharp. We see here the leading role of the US and the relatively high position of the OECD as a whole, even allowing for some members with weaker productivity levels. The five BRICS are a coming together of Brazil, Russia, India, China and South Africa. The BRICS countries have size – their combined population amounts to 42 per cent of the global population – but their average nominal GDP per capita is just under $7,000. They are also far from moving forwards at the same speed and their performance and level is dominated by China.

To crown our discussions in the previous chapters, we will now recalculate the size of the government sector in several alternative ways and compare these results with the official data. It will become apparent that the conventional measures are too narrowly defined. Consequently, if the methodological treatment of national accounts were changed, the findings of economic researchers would probably show different results on the relationship between government and other economic variables such as inflation, growth and unemployment.

Our analysis is, of course, limited by the current data availability. In spite of several data gaps, we have enough information to show that the officially presented figures constitute the lowest estimation of the size of government. Although our results are less favourable than those presented by national statistical offices, when it comes to redistribution or debt ratio, we believe that they provide a more realistic picture on both the size and the indebtedness of government.

Thus, we will use the aggregates recalculated in the previous chapter to present shares built on a different methodological approach. Of course, the way that relevant aggregates have been recalculated above might seem too radical, as well as the results. Nevertheless, as will be shown below, even conventionally quantified aggregates can be rearranged in a meaningful way, such that the size of government would be strikingly different than as officially presented. At the end of this chapter, we will also discuss the relation of statistics to relevant theoretical concepts.

GOVERNMENT EXPENDITURES RECALCULATED

Let’s start with the size of government in terms of total expenditures. Although the expenditures of public corporations are missing in our aggregates, we can still at least put the conventionally measured government sector into a different perspective by counting various government expenditures which are currently left out of macroeconomic models. Table 8.1 presents the results of the recalculation indicating the share between the recalculated government expenditure (see Table 7.2) and the recalculated GDP (see Table 6.5).

Table 8.1 illustrates the part of the recalculated GDP that is redistributed through the government accounts in the form of expenditure.

In this chapter, we shall make two steps beyond the conventional methodology. First, we shall extend the coverage of the government sector by including all public institutions as identified by the concept of control in national accounts; second, we will take into the accounts other kinds of expenditures which conventionally fall outside the definition. Government aggregates, recalculated in this way, will then be compared with nominal GDP as it is defined by current methodology.

Before going further, it is worth emphasizing that the analysis below is based, of course, on the currently existing data. Data availability poses, as usual for any empirical analysis, a more or less serious limitation. Simply put, not all useful and required data are reachable. For our purposes, this is mainly the case of public non-financial corporations and some of the central bank’s indicators. Unfortunately, lack of data does not allow us to recalculate all aggregates which would be desirable. Nevertheless, it opens a room for future research, once the missing data is available.

GOVERNMENT EXPENDITURES EXTENDED

First, let’s look at one of the most popular expressions of the size of government, in terms of total government expenditure. Although the designation “total” suggests that the government outlays are recorded in their entirety, this is not the case. Some categories of expenditures are excluded on the grounds that they will generate revenue in the future. In the national accounts terminology, these are categorized as “pure financial transactions”. Certain parts of government outlays thus remain hidden from the aggregate of “total expenditures”; technically speaking, they are identifiable only from the financial account in the system of national accounts, i.e. “below the line” and in the balance sheets. This concerns mainly financial investments in private or public companies through purchases of shares, provisions of loans and bond purchases as well as bond issuances which are not caused by a need to cover excessive expenditures. Despite the presumed recoverability, acquisitions of financial assets constitute, of course, a redistribution favouring one group of fellow citizens at the cost of others.

The statistical conception of market behaviour, as outlined in the previous chapter, obviously opens up many questions. The key one is whether market behaviour can be grasped realistically in a quantitative way. Market mechanism is actually a set of incentives; for its effective operation, the institutional environment matters more than anything else, which is why statisticians have focused their qualitative analysis at the level of competition and other institutional characteristics. However, here we face a serious limitation when it comes to measurability of many of the key features of market behaviour together with a rather arbitrary assessment of what is and what is not true market behaviour.

The issue of measurability is not new. It was Simon Kuznets who reminded us of the fact that measurable phenomenon can be captured statistically only if necessary information is available (Kuznets 1941). Kuznets’s observation on the measurability issue is a great point of departure. As we’ve indicated already the concept of control itself has serious weaknesses. Relying on the majority share is not sufficient, but reflection of partial ownership would make the compilation at the macroeconomic level impractical; moreover, the concept of excessive regulation remains open to its wider application. Nevertheless, not to overcomplicate our point, we will consider the concept of control as sufficiently broad to grasp the public sector in its entirety. This means that we will not question the size of the public sector, but we will focus on the line between the market and non-market sphere within the public sector.

We will begin with the concept of “the 50 per cent criteria” before proceeding to an analysis of the autonomy of decisions, the competitive environment and so on, i.e. the criteria which we term “qualitative”. We thus challenge the basic rules dividing the public sector between market and non-market parts. The central question is whether this approach is justifiable, i.e. whether in at least some cases a public units’ economic behaviour can be realistically likened to the behaviour of actual entrepreneurs, or whether the rules instead create a shadow area in measuring the size of government.

High-profile fiscal indicators are presented in relative terms, as a percentage of GDP. Although the general public is accustomed to this way of presentation, it is a mere convention. The truth is that GDP was not designed to serve this purpose. Rather it was meant to be an indicator of cyclical fluctuations and the statistical quantification of the taxable capacity of the economy for the purposes of public finance management during wartime. Since then the utilization of GDP has further expanded to such areas as well-being comparison, living standards measurement, etc. Along with these, the methodology has been developed into greater detail and it may have gone too far in its measurement of productive capacity of economies.

There is no denying that the relative size of government and government debt is a very comprehensible way to compare figures internationally and to show to voters the success or failure of economic policy. Nevertheless, from the methodological perspective, this approach does not make much sense. Even statisticians themselves admit that expressing the size of government in this way is not ideal and remains open to criticism (Lequiller & Blades 2014). One of the key points is what kind of information do we actually expect these figures to tell us? Is it the share of taxable income government drains from society? And does the relative size of government debt indicate whether government is capable of paying off its debt? Before answering these questions, we have to first discuss some details of the way GDP is calculated.

When showing the volatility in the statistical indicators and their methodology, we will seek to recalculate both the numerator and denominator of fiscal ratios. In doing so, we will exclude that part of the economy which does not, or cannot, serve as a source of government financing. It should be noted that GDP does not include, by definition, certain incomes which are commonly taxed, such as revenues from interest, rents on land and other natural resources and, not least, dividends. Of course, all calculations that follow are rough estimations based on simplistic assumptions, however, so are the official figures.

It was the nineteenth-century British prime minister Benjamin Disraeli who said “There are three kinds of falsehood, lies, damned lies and statistics”. This quip suggests that there are many contradictory stories that can be inferred from statistics and in the previous chapters we hope to have demonstrated that numbers on the size of government are no exception. This field of statistics suffers from many methodological drawbacks discussed above, but the way these figures are understood, interpreted, trusted and (over-)used by users is no less important.

The range of potential statistical outcomes, which depend on the methods applied and the techniques used, which was our aim to show, is something that statisticians do not reveal to end- users, but it is inherent to all statistical products. As we have shown, statistical methods can be applied in different, but meaningful ways, and input data can be rearranged, so that we arrive at very different descriptions of the financial position of governments, as well as the much wider statistical hold of government over the economy. Our arguments have implied that experts, analysts and the public put an excessive faith in the explanatory power of headline statistics and macroeconomic aggregates.

This uncertainty in the currently applied methods manifests itself in unstable and theoretically not well- anchored statistical rules which are still in a state of flux. One of the main issues we’ve discussed is the treatment of market behaviour which is currently not fully understood by compilers as a dynamic equilibrating process, but rather its analysis has been based on a static approach. This has consequences for all key aggregates like GDP, deficit and debt, leading to an underestimation of the statistical size of government. However, we do accept that the methodology has developed in the right direction over recent years. Indeed, the extension of the government sector to include a wider group of public institutions is highly desirable. Nevertheless, the methodology still has some way to go.

We argue that it would be more realistic to interpret officially published figures as the lowest estimation of the government’s influence over society.

We have finally arrived at a key point: the way government is currently portrayed in official statistics. Although our main concern is the delimitation of this particular sector in terms of institutions, we shall also address the issue of definition of individual aggregates such as total government revenues and expenditure, or even the definition of debt. Furthermore, to get a full picture, we shall also discuss some key techniques on how the compilation is based. These issues concern the consolidation practice and the choice between accrual or cash principle.

Concerning the last issue – the accounting principle – a brief comment is worth making. In general, all flows in national accounts should be recognized at the time the economic value is created, transformed, exchanged, transferred or extinguished (SNA 2.78); in other words, the accrual accounting principle shall be applied. This implies that the method of financial settlement does not matter. It sounds logical, but we should be cautious when interpreting macroeconomic aggregates that are compiled on this basis.

Consider a sharp growth in GDP; the economy is doing well optically, politicians are trying to take a credit for this, but it turns out that increasing insolvency is spreading over the economy. Technically, cash transactions might lag behind the trend in production (when customers pay late). If so, we can end up in a situation in which there might be widespread business bankruptcies, in contrast to the impression gained from statistical indicators.

Put differently, macroaggregates like GDP, GNI or government balance gives us absolutely no idea how their creation was financed. In these circumstances, it is quite easy for any government to inflate the growth by additional expenditure, not exceptionally financed by an excessive issuance of new government bonds. The construction of national accounts might be tricky for users in some cases, but to be aware of many peculiarities is decisive for understanding its explanatory power.

TAXES

Let’s start with taxes, i.e. compulsory payments paid to government for no direct consideration. Taxes constitute the predominant way of financing government activities. It mostly accounts for about 90 per cent of general government sector revenue; from the macroeconomic perspective, the share of compulsory payments in the EU accounted for nearly 40 per cent of GDP.

To understand the present, we shall make a short trip to the past. Luckily, we do not have to go too far back. At the time of writing there are still people living who can remember the advent of macroeconomics and macroeconomic statistics. Actually, societies had been functioning well before statistical institutes began to rattle off the tremendous volume of macroeconomic aggregates and all the related figures. How is it, all of a sudden, that we can hardly imagine living without these macroeconomic statistics? Indeed, do we really need this kind of information?

Our would-be need for macroeconomic statistics depends substantially on our economic position in the society today. Undoubtedly, members of a state-less society would see hardly any benefits in having macroaggregates at hand. But even now, in those societies that don’t come close to being “state-less”, most entrepreneurs hardly base their decisions on changes in GDP or in general price level; they would rather think of individual prices on local markets, they formulate their own expectations about future demand for individual products. Obviously, it also holds true for consumers who base their decisions on their personal incomes, family situation or subjective judgements. In their daily lives, individuals are not much interested in figures that tell them nothing about their individual situation or about the situation of their individual markets.

Nevertheless, we do not live in a state-less society. The state has become virtually omnipresent entering or aiming to enter into each and every economic relation on the market, aiming at stabilizing the economy and ensuring accessibility of certain goods and services as a reflection of its own policy. All this effort is translated into a bewildering patchwork of regulatory acts. To perform its agenda, the state needs to have statistical data. This is not surprising; the situation of politicians may be likened to any company’s owner who needs accounting information to manage its business. So do politicians. This brings us straight to the source of demand for macroeconomic statistics; it originates in politics and various state policies evolving over decades.

In 2016, the debt of Germany stood at 68.3 per cent of GDP with a downward trend. Among highly indebted European countries, Germany was considered one of the finest. However, what if we said that the debt of Germany stood actually at 314.9 per cent, with an upward trend? Would that be insanity? Not necessarily. Statistical reality is only a matter of statistical methodology and statistical theory. We, as data-users with critical thinking, should always ask ourselves a simple question to begin with: how can statisticians know that? We shall keep asking throughout this book: what methodology stands behind those figures? What are its weaknesses? And ultimately, are we really getting the information that we expected?

Questions of this kind might be asked in any statistical area. As Morgenstern (1963) rightly pointed out, statistical observations are deliberately designed, guided by constructed statistical theories which make macroeconomic aggregates graspable, but with aggregation comes costs. Errors of observations, calculation errors, misreporting and many other factors can introduce high uncertainty into final figures and their explanatory power. However, this typically remains hidden from data-users who get little idea of what degree of uncertainty is inherent in the final aggregates.

Let’s suppose that the most recent rate of GDP growth released by statisticians reached 2 per cent annually. During its calculation and the balancing process, a certain discrepancy between production and expenditure inevitably occurred. Statisticians then pick a method of how best to reconcile both sides of the balance. The final growth rate depends, of course, on their decision. They can increase total expenditure or to lower production, or vice versa, in all cases influencing the annual growth rate. It is entirely conceivable that the range of potential results obtained can vary from zero to 4 per cent growth. Therefore, we could have concluded that economic growth had come to a standstill, but we could have equally well concluded that we are experiencing solid economic growth.

As discussed in the previous chapter, macroaggregates provide guidance for many economic decisions of individuals and institutions. At least for this reason, we should take them seriously. In this chapter, we shall look at their use in economic policy. Discussing the influence of macroaggregates on economic policy is not only reasonable because the performance of economic policy is the main driving force of the demand for macroeconomic figures, but also because of the potential for economic-policy institutions guided by potentially ill-defined macroaggregates to negatively impact on society at large.

Politicians routinely use aggregates when arguing for the success of their economic policy or the failure of their opponents, as well as to emphasize a need for implementation of further measures should an “economic disaster” be avoided. Some have such faith in figures as the former French president Hollande who promised not to seek re-election in 2017 if unemployment continued to rise. That same year, newly elected president of the United States Donald Trump promised to return to 4 per cent annual growth. Figures simply matter.

Similarly, government deficit and debt figures have had significant political impact in many European countries where they have gained a prominent position chiefly in connection with the political project of European Monetary Union (EMU). Due to the interlinking between fiscal and monetary conditions, an inherent tendency of governments to run deficits and to amount debts may pose a problem especially for the functioning of the EMU. To enforce fiscal discipline and to avoid a collapse of the eurozone, which is based on the fiat currency, the EU bureaucracy developed a comprehensive and highly complicated statistical system known as Excessive Deficit Procedure (EDP). In addition, they put these figures at the centre of fiscal management.

MACROAGGREGATES AND THE EUROPEAN UNION

Why is this EDP “monitoring system” required? The main reason is a concern that unsound government finance can prevent the central bank from delivering price stability. If any member of the monetary area runs an unsustainable level of debt, than bailout could be necessary.

Let’s start with the question, why the discussion on the measurement of government should be important at all. Following fundamental changes in the field of mainstream economic theory in the first half of the twentieth century, the role of government in economic affairs has become one of the predominant concerns of economists. So, economic researchers have put a great deal of effort into finding out an optimal size or scope of government, which draws on mutual causal links between relevant macroeconomic aggregates like government expenditures and economic growth, or government indebtedness and inflation rates. Undoubtedly, statistics, that is the quantitative measurement of abstract economic concepts, is the sine qua non of empirical research.

To bring economic concepts into existence by way of appropriate quantifications is left almost exclusively to statisticians. Admittedly, after many decades of rapid development in economics and statistics, we have arrived at the point where economists rely on statistical quantifications, but are at the same time little aware of how these figures have been compiled. On the other hand, most statisticians have little idea of all the purposes for which the result of their work is used. This situation is apparently further enhanced by a precipitous development on both sides, i.e. a growing complexity of the methodology of macroeconomic data and a broadening of its use in economic policy.

In the case of macroeconomic aggregates, we are not talking about monetary categories easily observable. As Morgenstern (1963) reiterates, statistical observations are based on the constructed, deliberately designed and invented theory. The compilation carried out by professional statisticians is a lengthy and highly complicated process. Lack of information and methodological complexity creates an environment in which an occurrence of measurement errors is inevitable. In his famous book, Morgenstern (1963) further enumerates several sources of error negatively affecting the accuracy of economic calculations. For the sake of our argument, chiefly the lack of clear definitions of the government sector, one of the sources of errors stated by Morgenstern, is of vital importance. Morgenstern rightly notes that weak and unsteady definitions cannot provide a stable ground for reliable quantifications.