Scoreboard for the Surveillance of Macroeconomic Imbalances in the European Union

Abstract

This chapter focuses on tools to counter macro-economic imbalances in the EU. The economic and financial crises have revealed the limits of the original EMU surveillance tools. To prevent and correct macro-economic balances, the Macroeconomic Imbalance Procedure (MIP) was adopted. In this chapter, the authors present an overview of the MIP, its Alert Mechanism, and the rationale behind the indicators. The scoreboard is included in the yearly Alert Mechanism Report (AMR), which serves, inter alia, as a basis for the Commission to draw a list of Member States that are subject for an in-depth review. The authors notice that the number of countries that are in the list has not decreased over the years because of external sustainability issues, export performance and competitiveness, private sector indebtedness, and housing and asset markets issues. Yet, there have been improvements in terms of deficits, gains of price and non-price competitiveness, and the export performance is more homogeneous. Private debt stocks persist and unemployment seems to have increased.

Annexes

1.1 Annex A.2.1: The Scoreboard Indicators in More Detail

This annex looks more in detail at the scoreboard indicators, their defnitions and also reports some of the concerns at the time they were selected, including alternative indicators considered. In several instances, a number of indicators/transformations were considered, each having particular strengths and weaknesses. After careful consideration of the pros and cons of these alternative options, the most appropriate indicator was chosen. Nevertheless, recognizing the critical importance of taking due account of country-specific circumstances and institutions, the economic reading of the scoreboard is complemented by additional information and indicators. This inter alia includes the general macroeconomic situation, such as growth and employment developments, nominal and real convergence inside and outside the euro area and specificities of catching-up economies. Additional indicators are considered that reflect the potential for the emergence of imbalances as well as the adjustment capacity of an economy, including its potential to sustain sound and balanced growth, such as different measures of productivity, inflows of FDI, capacity to innovate and energy dependence. The state of financial markets, which played an important role in the current crisis, will also be covered.

1.1.1 The Current Account Balance Indicator

1.1.1.1 Definition and Data Sources

The scoreboard indicator is the 3-year backward moving average of the current account balance expressed in percent of GDP, based on Eurostat data from Balance of Payments statistics, with the indicative thresholds of +6% and −4%. The average over 3 years is used so as to control for short-term fluctuations of the annual figures and to provide indications of the persistence of a potential imbalance.

Data on the current account balance²¹ are derived from the Balance of Payments (BoP) statistics reported by Eurostat. This source is widely used by other international institutions as well as academics. BoP (and International Investment Position) statistics are the statistical tools expressly built to monitor the relations of a country with the rest of the world. An important advantage of this data source is also its quick availability and high frequency. BoP data also allow decomposing external imbalances by counterpart area, hence giving an idea of a possible spill-over of a crisis from a given country to another.

An alternative data source on current transactions balances is the Rest-of-the-World Accounts (RoW) in the National Accounts (NA). This data is consistent with other indicators derived within the NA framework and also with the Commission forecast for the current account balances. However, there are discrepancies between the data derived from the NA and the BoP data. These differences occur despite the fact that “current external balances” from the NA and “current account balances” from the BoP describe the same economic concept. The issue has been closely monitored by Eurostat together with ECB and national statistical institutes and it appeared that the differences stem from compilation practices, methodological reasons, different data vintages and revisions, errors and omissions²². BoP data are compiled first, and subsequently incorporated in relevant external account components of NAs. When compiling NAs, data related to the RoW sector have to be reconciled with those related to the domestic economy (the focus for NAs).

1.1.1.2 Indicative Threshold

A simple statistical distribution analysis provides an indicative threshold for current account deficits of −4%. This indicative threshold was derived from the data sample starting in 1970 for most of the old Member States and in early/mid 1990s for the new Member States, and ending in 2007. It appears reasonable to compute the value of the threshold on the basis of a long period which extends beyond the last decade characterized by increasing divergences in the euro area. The increase in the divergence of external positions in the EU over the past decade together with the inclusion of new Member States with typically high current account deficits would introduce a downward bias in the sample.

This threshold value is also broadly in line with the evidence from the empirical literature on balance of payment crises and sustainability of current account imbalances. There are broadly three strands of this literature, which are relevant for the determination of the threshold²³: Firstly, a number of research papers investigate past episodes of significant current account adjustments and attempt to identify some regularities, including the levels of current account deficits at which the adjustment starts. Examinations of past episodes of current account adjustments show that a typical current account reversal starts at around −5% of GDP (Summers 1996). Freund (2005) found on a sample of industrialized countries that the mean for the current account to GDP ratio at the beginning of large current account adjustments was around −6.3% (median was −4.9%). Similarly, IMF (2007) found on average that past current account reversals in advanced countries started when the current account deficit stood at about 4.1% of GDP. Reversals of persistent current account surpluses typically started at the level of 2.4% of GDP. The corresponding values for an EU sub-sample would be −4.3% and 2.5%, respectively²⁴. The results of all these studies, nevertheless, show that there is a very significant variance across countries and the thresholds should be interpreted with caution. Using an alternative approach to examining the determinants of past recessions (binary recursive trees), Ghosh and Ghosh (2003) find that countries with current account deficits above 2.5% of GDP have a sevenfold greater probability of a crisis than countries with smaller deficits;

Secondly, current account norms, i.e. current account to GDP ratios as justified by fundamentals are usually computed based on a reduced form of a panel econometric model in the spirit of Chinn and Prasad (2003). The results have to be interpreted with utmost caution as they are subject to numerous conceptual and methodological caveats. Tentative estimations of current account norms for the EU indicate that the average current account deficit should be around −4.7% of GDP (median −3.4% of GDP) and the average “justified” surplus around 3.7% of GDP (median 3.1% of GDP).

Finally, much research has focused on assessing the sustainability of current account imbalances. This strand of literature typically attempts to estimate values of current accounts which would stabilize the external position of a country at the current or a predetermined level (e.g. Milesi-Ferretti and Razin 1996; Edwards 2001). These results are typically country-specific and do not deliver a general benchmark.

The upper value of the threshold is set at +6%. The upper quartile of the distribution of the 3-year backward average of current account balances corresponds to +2%. To this an additional 4% margin has been added in line with the “intelligent symmetry” approach to current account balances. This allows tackling both current account surpluses and deficits but recognizes that the urgency for policy intervention is clearly greater in the case of current account deficits. It also reflects the fact that the risk of negative spillover effects of current account deficits is more prevalent than for current account surpluses due to sustainability considerations.

1.1.1.3 Complementary Indicators

In the discussions it was also agreed that it is important that the economic interpretation will take due account of additional relevant information, in particular the specificities of catching up economies. The potential risks from external deficits need to be qualified by taking into account capital transfers in the form of EU structural funds, as they can finance in part current account deficits. Similarly, the destination of the capital flows is relevant as strong FDI inflows help to provide a relatively safe financing of current account deficits in many of these Member States. To account for the inflows of EU structural funds, the sum of current account and capital account will be considered for Member States for which this information is relevant. Conceptually, the sum of current account and capital account determines the net lending/borrowing of a country and is thus the flow counterpart of the net foreign financial asset position/net international investment position. The capital account comprises (a) capital transfers receivable and payable between residents and non-residents (e.g. debt forgiveness), and (b) the acquisition and disposal of non-produced, nonfinancial assets between residents and non-residents (e.g. natural resources, licenses, contracts, leases or marketing assets). The net size of the capital account is typically rather small. However, in a number of catching up Member States, capital account can be non-negligible as a part of structural/cohesion funds is recorded here.

1.1.2 The Net International Investment Position

1.1.2.1 Definition and Data Sources

The scoreboard indicator is the net international investment position expressed in percent of GDP based on Eurostat data from Balance of Payments statistics, with the indicative threshold of −35%. This indicator is calculated as a share of GDP to allow for cross-country comparability. As this is a stock indicator, the value for the last available year is used.

For consistency reasons, data on the NIIP²⁵ are derived from the Balance of Payments statistics reported by Eurostat, i.e. the same data source used for the current account balance. Like in the case of current account balance, there is an alternative data source – the Rest-of-the-World Accounts (RoW) in the National Accounts (NA). The general considerations entering the selection of the data source are essentially identical to those concerning the indicator on current account balance. In this case, the differences between the two data sources are considerably larger than for current account data. In addition, while Eurostat has extensively analysed the discrepancies between BoP and RoW (NAs) in the current and capital account, little is known about the discrepancies observed between national IIP and NFAs (RoW) data.

1.1.2.2 Indicative Threshold

The statistical analysis of the NIIP distribution yields −35% of GDP as an indicative threshold. It is difficult to establish a level of net external assets which can be considered as risky and economic literature attempting to do this is rather scarce. This is due to the fact that next to the absolute level of net foreign liabilities, it is in particular the composition of both gross assets and liabilities in terms of types or maturities, which determine the overall vulnerability of the external position of a country.

Unlike large negative NIIP positions, large positive external asset positions are not a priori considered to be problematic for a Member State or the functioning of EMU. Therefore, the scoreboard contains an indicative threshold for negative NIIP only.

1.1.2.3 Complementary Indicators

NIIP is a good starting point in the assessment of external positions of Member States. However, the composition of NIIP is important for a deeper understanding of the degree of vulnerability of a country. Therefore, also in this case, the economic reading of the scoreboard will take account of additional relevant information.

In this sense, it is useful to focus specifically on liabilities that require repayment of principal or interest, separately from non-debt generating liabilities. This provides useful additional information to interpret the overall NIIP as these components have an impact on external solvency of an economy. This distinction is important especially for the specificities of external positions of catching up Member States, which experience strong Foreign Direct Investment (FDI) inflows. It can be argued that FDI constitutes a relatively less risky and more stable form of financing than other alternatives and thus these inflows do not increase country’s vulnerability to the same extent²⁶.

In this respect, the economic interpretation will consider the indicator on Net External Debt (NED), which, compared to the NIIP, does not contain portfolio FDI²⁷, portfolio equity and financial derivatives. By focusing on external debt liabilities, i.e. those that require payments of principal and/or interest, NED further qualifies the assessment of the riskiness of a country’s external asset position²⁸.

1.1.3 The Real Effective Exchange Rate

1.1.3.1 Definition and Data Sources

The scoreboard indicator on REER is calculated as the 3-year percentage change of the nominal effective exchange rate (NEER) deflated by the consumer price index deflators, relative to a set of 41 industrial countries, with DG ECFIN as the data source and with the indicative thresholds of +/−5% and +/−11% for euro-area and non-euro-area countries, respectively^{29 , 30}.

In order to derive the REER from the NEER, several options were discussed during the design of the scoreboard. The competitiveness of each supplier relative to its trading partners can be measured by the REER expressed either in terms of production costs (ULC), export prices or economy-wide prices (HICP or GDP deflators).

First, the REER based on broad measures of prices or costs, such as HICP or GDP deflators, provides the most comprehensive picture of price competitiveness of domestic producers in a medium-term perspective. The basket of goods on which these price indexes are calculated includes both tradable and non-tradable goods (excluding capital goods). Additionally, given that price indexes also include the price of imported goods, countries with different import-dependency will have different relative price effects of nominal exchange rates changes. Such effects need to be accounted for when interpreting the REERs. Second, the ULC-based REER shifts the focus of the assessment of relative competitiveness in terms of consumer prices to relative production costs. This important notion is also picked up by the ULC scoreboard indicator (see below). For tightly integrated economies in a monetary union, a ULC-based REER would capture a similar notion as the headline ULC indicator. Third, the REER based on relative export prices, while being a rather intuitive measure of market competitiveness, suffers from a number of weaknesses; (i) the calculation of export prices is strongly influenced by the composition of exports and by the price dynamics of exported goods; (ii) REERs based on export prices convey information on how producers set prices in order to maintain market shares in case of nominal exchange rate variation (pricing to market) even at the expense of profits, providing a short-term picture that might be out of line with the dynamics of REERs calculated with different deflators.³¹

1.1.3.2 Indicative Threshold

Concerning the indicative thresholds, symmetric thresholds are considered for the REER indicator. The focus is put on detecting harmful imbalances, which may be captured by an unsustainable appreciation meaning a loss of competitiveness, or depreciation signaling potential problems related to domestic demand or the potential of harmful future price convergence. Furthermore, a differentiation of thresholds between euro-area and non-euro-area countries is adopted in line with the Herman Van Rompuy Task Force (Task Force 2010). Differentiated thresholds reflect nominal exchange rate variability, catering thus for that countries with flexible exchange rates may be subject to non-persistent swings in the REER due to nominal exchange rate fluctuations with their most important trading partners.

The differentiation between euro-area and non-euro-area Member States also reflects the trend real appreciation in catching-up countries. This can be explained by increases in wages in the tradable sector due to productivity growth that are transferred to the wages and prices of the non-tradable sector (Balassa-Samuelson effect) where productivity does not increase commensurately. Countries that have undergone economic transitions (e.g. liberalized trade and capital flows), and have been catching-up to the levels of development of the EU-15 countries, typically have experienced a trend appreciation in terms of the REER indicator. If REER appreciation is due to the Balassa-Samuelson effect, with productivity improvements in tradable goods, this should not threaten international competitiveness. The most recent empirical studies find a Balassa-Samuelson effect for new Member States of only 1% per year, on average (Égert et al. 2005). This is a rather modest contribution that is not sufficient to explain the observed REER appreciations in catching-up countries.

Overall, with a REER indicator calculated as a 3-year percentage change, the transformation looks at medium-term developments in relative prices. To also cater for exchange rate flexibility, one standard deviation is added to the value of the thresholds derived from the distribution in the sample of euro-area countries. The standard deviation is larger than the value on the Balassa-Samuelson effect estimated in the literature, i.e. 1% change per year as signaled above. The thresholds corresponding to the lower and upper quartiles of the distribution are −/+5% for the 3-year percentage change. These thresholds would apply to euro-area countries³². For the non-euro area countries, the standard deviation of the distribution is subtracted from the lower quartile and added to the upper quartile. The resulting thresholds for non-euro-area countries are therefore −/+11%.

1.1.3.3 Complementary Indicators

The REER indicator captures persistent price changes in a common reference unit (HICP/CPI) relative to major trading partners and thus illustrates the magnitude of developments in price and cost competitiveness. Significant deviations of the REER based on HICP/CPI from the benchmark indicate that prices have grown out of line with productivity for some time without compensation via the nominal exchange rate, i.e. the country has lost or gained labor cost competitiveness with respect to its trading partners.

In particular for euro-area Member States, persistent divergence in price and cost competitiveness versus their EMU peers is a concern as this may hamper the smooth functioning of the monetary union. In order to monitor such structural losses or gains in competitiveness and trade, the additional indicators complement the economic reading with a REER indicator that focuses on euro area trading partners instead of the broader set of 36 countries in the headline REER indicator. Moreover, REER developments are analyzed in conjunction with other scoreboard indicators on competitiveness (in particular the development of ULC and export market shares) to gain insight on the cost, price and non-price competitiveness performance of Member States.

1.1.4 Export Market Shares

1.1.4.1 Definition and Data Sources

The scoreboard indicator is the percentage change of export market shares over 5 years, based on Balance of Payments Eurostat data, with a lower indicative threshold of −6%. For each country, the export market shares are computed as the share of the country’s export revenues in total world export revenues, in current prices. The indicator thus adds many aspects of competitiveness to the scoreboard that are not captured by price and cost competitiveness alone (that is monitored with the real effective exchange rate based on HICP/CPI and the nominal ULC).

There are a number of options available as regards the definition of the indicator. Firstly, one aspect to take into account is the time variation to apply: changes over one, 3 or 5 years. Given the high volatility of year-on-year changes in view of idiosyncratic trade shocks, this option was excluded in favor of a longer assessment period which would better reflect structural losses/gains in export performance. The percentage change over 5 years of the value of goods and services exports for each country as share of the world exports of goods and services appears to be the most opportune data transformation to measure long-term competitiveness development. There is an important caveat, though: the short time series available permits to calculate 5-year export market shares changes only from 1999 onwards.

1.1.4.2 Indicative Threshold

The indicative threshold of the export market share indicator has been obtained from the lower quartile of the data series distribution. This threshold corresponds to cumulative losses of 6% over a period of 5 years. For this indicator, no upper threshold has been considered because in the context of the MIP, since the focus is on the detection of the harmful imbalances that may jeopardize the healthy functioning of the EMU. In that context, the key concern is the detection of Member States with deteriorating competiveness positions given by unsustainable losses in export market shares.

1.1.4.3 Complementary Indicators

The economic interpretation of the export market shares indicator is performed in conjunction with other long-run scoreboard indicators. In fact, most of the fluctuations and country differences in current accounts are driven by developments in the balance of goods and services, which is usually the largest component of the current account. Losses in competitiveness, the build-up of large current account deficits and the deterioration of the net international position in some Member States can be related to a range of underlying domestic macroeconomic imbalances.

Export market shares could also be computed with trade data in volumes (at constant prices) rather than with data at current prices (using Balance of Payments data on exports) so to avoid biases deriving from relative price developments. Such an indicator has the advantage to exclude variations that are due to relative export price developments. While the indicator calculated at current prices covers data on goods and services, the variation of export market shares in volumes only covers exports of goods, given the lack of reliable deflators for trade in services. The current price data series for goods and services has therefore been chosen as indicator in the scoreboard for coverage reasons, while export market shares (for goods) in volumes will complement its reading among the additional indicators.³³

Furthermore, with respect to ‘non-price’ competitiveness, the scoreboard already includes several indicators that are directly or indirectly related to competitiveness at large, i.e. the change of the REER based on the HICP/CPI deflators, the change in export market shares and the change of ULC. Hence, the dynamics of ‘price’ and ‘cost’ competitiveness together with the variation of export market shares offer an indication of ‘non-price’ competitiveness which in turn can be defined as the “export performance that cannot be explained by price developments”.³⁴ In order to gain more precise insight into such developments, the reading of the scoreboard also relies on value-added decompositions and analysis according to sectoral export market shares.

In addition, and as highlighted by the so called “new-new trade theory”, in the long run the driver of exports is productivity (Melitz 2003). Only the most productive firms in each country export, and countries’ export performance is closely related to changes in average productivity.³⁵ Therefore by including export market shares, the scoreboard includes not only ‘non-price’ competitiveness elements but also, indirectly, productivity. In order to disentangle this feature, the scoreboard includes a productivity indicator among the additional indicators used for the economic reading. This indicator is measured as the year-on-year growth of labor productivity expressed as GDP per person employed (at constant prices, 2000). Taking account of productivity developments, in particular during protracted periods of low growth, is relevant as macroeconomic imbalances are often symptomatic for a lack of productive investments. Indicators of productivity growth are thus not read as a direct early warning indicator for emerging imbalances, but used in conjunction with forward-looking scoreboard indicators in order to obtain a better understanding of the potential severity of macroeconomic imbalances (in terms of their likely persistence and the capacity of the economy to adjust).

Finally, the rise of emergent countries in the world trade impacts all EU members and all advanced economies suffer losses as the world trade structure is changing. The indicator of world market shares does not disentangle losses in market shares that are specific to each country and those that concern all advanced economies. To better understand the causes behind the losses in export market shares, an auxiliary indicator compares the export performance of each country with the export performance of the OECD economies.

1.1.5 Unit Labor Costs

1.1.5.1 Definition and Data Sources

The scoreboard indicator is the percentage change over 3 years of nominal unit labor cost based on Eurostat data, with the indicative thresholds of +9% and +12% for euro-area and non-euro-area countries, respectively.

The ULC index used in the scoreboard corresponds to the ratio of compensation per employee to real GDP per person employed (labor productivity). The original data on nominal compensation per employee, GDP and employment stem from Eurostat and the index is calculated by DG ECFIN (AMECO database).³⁶ In order to capture the medium/long term developments of labor costs, the scoreboard indicator for the ULC is calculated as the 3-year percentage change, as it dampens cyclical impacts on this indicator and keeps memory of built-up competitiveness losses. Besides the percentage change over 3 years, the year-on-year percentage change of the ULC index and the deviation from the long term average were also computed. Nonetheless, these latter transformations are either too volatile or heavily influenced by the average trend in ULC in each country.

1.1.5.2 Indicative Threshold

The threshold corresponding to the upper quartile of the statistical distribution over the sample of euro-area countries is 9%.³⁷ For non-euro-area countries 3 percentage points have been added to obtain a 12% threshold. This differentiation is not based on the statistical distribution over the non-euro area sample, but was made since historical data reflect the fact that the majority of non-euro area countries have experienced a major trade liberalization in the period covered by our data (since 1995), which entails a natural process of factor price equalization towards the levels of the trade partners. These strong adjustment processes due to trade liberalization should however be considered to weaken over time and in the future. In that respect, catching up transition economies are of particular concern as they can experience a trend increase in ULC because the increases in wages in the tradable sector linked to productivity growth are transferred to the wages and prices of the non-tradable sector (Balassa-Samuelson effect), where productivity does not necessarily increase. However, recent empirical studies gauge this effect to be limited (Égert et al. 2005; European Commission 2008; and Peters 2010). No upper threshold has been considered, because in the context of the MIP, the focus is on the detection of the harmful imbalances that may jeopardize the smooth functioning of the EMU, such as unsustainable increases in the cost of labor.

1.1.5.3 Complementary Indicators

The interpretation of the medium/long-run ULC indicator will be complemented with the scoreboard indicators on competitiveness and trade. The ULC indicator together with the (HICP-based) REER indicator allows a comprehensive assessment of the cost/price competitiveness developments in each Member State. Large and sustained increases in ULCs may lead to the erosion of competitiveness, especially if combined with a widening current account deficit and declining market shares for exports. For instance, in the years preceding the present crisis, wage growth outstripped productivity improvements in many Member States, inducing sharp increases in ULC. Similarly, the developments in REER, which show price and cost competitiveness relative to the main trading partners, point to increased divergence.³⁸ This may signal potential structural rigidities in product and labor markets but partly reflects the catching-up process in several Member States. To account for the longer-term losses in cost competitiveness, percent variations over longer time periods (up to 10 years) are also considered in the economic reading of the scoreboard.

Moreover, scoreboard indicators on competitiveness and trade are complemented by a set of additional indicators. Persistent divergence in price and cost competitiveness among euro-area countries is of particular concern, provided that ensuing external debt problems may hamper the smooth functioning of EMU. For this reason, an effective ULC deflator³⁹ indicator versus the rest of the euro-area countries is included among the additional indicators. The importance of monitoring effective ULC deflators as an indicator of competitiveness developments was recognized by the Heads of State or Government of the euro area in their Council Conclusions of 11 March 2011. As part of efforts to assess whether wages are evolving in line with productivity developments, the Pact for the Euro Area calls for ULCs to be monitored over a period of time, by comparing developments in other euro-area countries and their main trading partners.

1.1.6 House Price Index

1.1.6.1 Definition and Data Sources

The scoreboard indicator is the year-on-year growth rate of the deflated house price index (HPI)⁴⁰, data source Eurostat, with an indicative threshold of 6%. The consumption deflator is used to reflect the value of house prices relative to the whole consumption basket. This way of computing real house prices is widely used in the literature and by other international organizations (e.g. OECD).

Data on house price indices are provided by various institutions. The only harmonized index is, however, the Eurostat experimental house price index (HPI).⁴¹ It aims at measuring price developments of all residential properties purchased by households (flats, detached houses, terraced houses, etc.), both new and existing, independently of their final use and their previous owners. Only market prices are considered (mirroring the practice of the HICP), self-built dwellings are therefore excluded. The land component is included in the HPI.

Since 2005, Eurostat has been collecting HPI data from several Member States in the framework of the Owner Occupied Housing project. When the process of selecting the scoreboard indicators started, HPI data were available for 17 Member States for the period 2005-Q1 to 2010-Q1. As of the data extraction date for the Alert Mechanism Report⁴², important progress has been made and data have been collected for all 27 Member States, at least for 2010 and 2011. The publication of HPI quarterly data as from 2005 started in December 2010 and the most recent release was in January 2012. As of 2013, Eurostat officially launched its Property Price Index, derived from its initial experimental indicator (see Eurostat 2013 for more details). In the medium-run, Eurostat will work on providing longer time series for the HPI, starting possibly in the mid-1990s.

For time series analyses, a longer time sample is needed. To this end, other data sources such as the ECB and the OECD could be used, given that, for the period 2005–2010, the correlation between the growth rates of Eurostat HPI and of the ECB and OECD house price indicators is very high for a large majority of Member States. However, for some new EU non-euro-area Member States, only the data from the Bank for International Settlements database (BIS) are available. BIS data are the least harmonized, as they use a variety of prices, such as price per square meter, per standard flat, etc.

1.1.6.2 Indicative Threshold

Given the scarcity of time series data, it is difficult to derive a threshold based on the statistical distribution. Using the OECD dataset of 19 OECD countries on a long series of historical data (1970–2007) gives a lower upper quartile of the distribution of 6%. This compares with the threshold derived from the information provided by the house price cycle. For instance, a recent study by Agnello and Schuknecht (2009) looks into house price cycles and identifies phases of booms and busts in 18 industrialized countries. The 25 most severe booms are characterized by an average expansion of real house prices of 40% over an average period of 7 years. (The severity is judged based on an index which gives an equal weight to the magnitude and the duration of the house price in the boom phase.) This translates into an annual increase of close to 6%. Given that only the top 25 most severe booms over the period 1970–2007 are selected amongst the total of 100 identified booms, the associated 6% threshold could be seen to be at the high end.

1.1.6.3 Complementary Indicators

As part of the economic reading of this scoreboard indicator, real house price growth over longer time periods will also be considered, as a complement to the short-run indicator. To this end, 3-year average price growth rates are used as an additional indicator. Moreover, the analysis of the house price cycle proves to be very informative. The cumulated house price growth from the latest through to peak and the average annual pace of growth can illustrate the scale of house price developments. Coupled with information on house price determinants, such as credit growth, cost of credit and demographic factors, these could provide indications of future house price developments.

During the process of designing the scoreboard, the nominal house price index was also discussed; this indicator is likewise used in the economic reading. For instance, if nominal house price inflation occurs at the time of final consumption inflation, and thereby the real house price growth does not pick up the acceleration in nominal house price inflation, potential risks of a house price bubble will be grasped through economic judgment. In order to put house prices into perspective, it is useful to assess them against households’ capacity to repay and alternative options such as rental markets. In this vein, affordability (price-to-per capita disposable income) and dividend (price-to-rent) ratios will also be assessed. Although their findings have to be considered with caution due to their simplifying assumptions and their crude approach, they provide a useful qualifier.

Volume indicators, in particular residential construction and value-added in construction (as percent of GDP), are a useful complement to assess house prices. The responsiveness of supply to changes in prices plays an important role in shaping housing markets. A responsive housing supply reduces house price volatility but at the potential cost of greater fluctuations in residential investment, with the net impact on overall economic activity being unclear (Andrews et al. 2011). Thus, it seems that during boom periods, inelastic housing supply reinforces house price overvaluation while high supply elasticity coupled with expectations of future housing price rises may lead to overshooting in construction activity.

1.1.7 Private Sector Debt

1.1.7.1 Definition and Data Sources

The scoreboard indicator is the stock of private sector⁴³ debt in% of GDP, defined as the sum of loans and securities other than shares, consolidated. The threshold of private sector debt is 133%.

Private sector debt is a stock variable defined as the sum of loans and securities other than shares (excluding financial derivatives)⁴⁴ and is expressed in percentage of GDP. The data stem from the annual financial accounts and balance sheets (AFA) collected by Eurostat and the quarterly financial accounts (QFA) collected by the ECB.⁴⁵

The envisaged indicator is currently based on consolidated data, i.e. excluding intra-sector liabilities such as intra-enterprise loans. When the scoreboard was initially designed (European Commission 2012), however, it was decided to use non-consolidated data mainly because of lack of consolidated data for all Member States. One drawback of non-consolidated data is that it is not known to which extent intra-sector liabilities are dominated by intra-group transactions. If intra-group loans form the bulk of intra-sector credit, non-consolidated data may be biased due to national and multinational (non-financial) corporate accounting practices. For example, in Member States where each unit/branch of an enterprise-group reports on its credit/debt, the non-consolidated data would probably show higher figures than in Member States where the headquarter reports on total group consolidated credit/debt. Thanks to the technical work by Eurostat and the Member States’ statistical institutes, consolidated data are now available for all Member States and used as sources for the headline scoreboard indicator.

1.1.7.2 Indicative Threshold

The threshold of private sector debt is 133% of GDP, as derived from the upper quartile of the statistical distribution of the indicator. Annual data for the period 1995–2007 were used to establish the threshold.

1.1.7.3 Economic Interpretation

The selection process of the indicator dismisses the category “other accounts: payable”. Although it is a non-negligible subcategory for several Member States, it exhibits high volatility and may therefore introduce noise in the data that is difficult to justify. The item reflects valuation effects as well as volume effects (mainly reclassifications), but the two are difficult to disentangle. Consideration is to be given in the economic reading to the third subcategory mentioned above: “other accounts: payable”. Other payable includes: “trade credits” and “other payable excluding trade credits and advances”. The latter consists of financial claims which arise from timing differences between distributive transactions or financial transactions on the secondary market and the corresponding payment, for example: (a) taxes; (b) social contributions; (c) wages and salaries; (d) rents on land and subsoil assets; (e) dividends; (f) interest; (g) transactions in financial assets on the secondary market.

In order to assess how consolidation practices compare across Member States. Meanwhile, non-consolidated data will be used as an additional reading indicator.⁴⁶ By including intra-sector debt, the use of non-consolidated data acknowledges that apart from bank loans, an increasingly important source of financing may be intra-sector. When large differences between consolidated and non-consolidated data exist, the Commission services will examine the reasons behind. The Commission will examine, jointly with Eurostat, whether intra-enterprise loans dominate intra-group liabilities for non-financial corporations or whether there are other reasons in order to shed more light on the consolidation practices across Member States.

Moreover, Monetary and Financial Institutions (MFI) data on loans, collected by the ECB, will also be considered as part of the subsequent economic analysis. The advantage of using MFI loans consists in their widely spread use, both by academics and international organizations. As a disadvantage, securities which are also a source of financing for non-financial corporations, are not included, overlooking thus country heterogeneity with respect to firm liabilities’ structure. Also intra-sector credit, which may be an increasingly important source of financing, is not captured when using MFI loans data.

1.1.8 Private Sector Credit Flow

1.1.8.1 Definition and Data Sources

The scoreboard indicator is private sector credit flows (transactions) expressed in% of GDP, and it includes loans and securities other than shares (excluding financial derivatives), consolidated data. It is the flow counterpart of private sector debt (which is a stock indicator). The indicative threshold of private sector credit is 14%.

The sources of data are the annual financial accounts and balance sheets (AFA) collected by Eurostat and the quarterly financial accounts (QFA) collected by the ECB. The source data used for debt and credit flows is the same. Therefore, data, methodological and technical issues pertaining to these two indicators largely overlap.

Two other indicators were considered and discarded. Firstly, initial considerations aimed at an indicator measuring the year-on-year percentage change in credit flow. The rationale behind this choice of data transformation was that it can detect rapid increases in credit flows that could be associated with credit bubbles, which in turn may contribute to crisis situations. However, interpretation difficulties arise since credit flows typically evolve in a cycle. This induces a risk that by using this indicator the gradual build-up of a credit bubble is concealed when credit flows remain high but steady (“high speed but no acceleration”) and thus its early-warning properties are jeopardized. Secondly, the year-on-year change in private sector debt as percent of GDP was considered, as it represents the most straightforward flow counterpart of the indicator on private sector debt. Notwithstanding its consistency with the stock variable, this indicator is heavily influenced by Other Economic Flows (OEF), which is a non-directly interpretable residual. OEF consists of nominal holding gains and losses (changes in prices) and other changes in volume (mainly reclassifications). However, distinguishing between changes in prices and changes in volumes is difficult, and it seems that OEF is heavily influenced by reclassifications.

1.1.8.2 Indicative Threshold

The indicative threshold of private sector credit is 15% of GDP, as derived from the upper quartile of its historical distribution. Annual data for the period 1995–2007 are used to establish the value of the threshold.

1.1.8.3 Complementary Indicators

As in the case of private sector debt, the subcategory “other accounts: payable” is not included, although this item is potentially interesting to be considered as an additional indicator to qualify debt developments. Moreover, as discussed for the private debt indicator, an important issue is the choice between consolidated or non-consolidated data for the scoreboard indicator. In order to ensure consistency with the stock counterpart of credit flows, the latter is also based on consolidated data, i.e. excluding intra-sector liabilities. Non-consolidated data will be used as an additional reading indicator.

1.1.9 General Government Debt

1.1.9.1 Definition and Data Sources

The scoreboard indicator is general government debt in percent of GDP, defined under the Excessive Deficit Procedure (EDP) as the total gross debt at nominal value outstanding at the end of the year and consolidated between and within the sectors of general government. The threshold is 60%.

The definition of general government consolidated gross debt is the one used for the purpose of the Excessive Deficit Procedure (EDP) as well as for the Stability and Growth and Stability Pact (SGP). The Maastricht Treaty, together with Council Regulation (EC) No 3605/93 define the general government debt as the total gross debt at nominal value outstanding at the end of the year and consolidated between and within the sectors of general government. Other accounts payable and financial derivatives are not included in the definition, mainly for measurement reasons.

1.1.9.2 Indicative Threshold

As regards the threshold for the general government’s indebtedness, the Treaty reference value of 60% of GDP will be used (as a separate indicative threshold for public debt under the MIP would be confusing).

1.1.9.3 Complementary Indicators

General government debt is assessed for its contribution to the general indebtedness of a Member State, being thus looked at together with private sector debt. A high level of general government debt is more worrying when it accompanies large private sector debt. Nevertheless, high general government debt represents a vulnerability per se. A high level of government sector debt cannot in any way compensate for a low level of the non-financial private sector debt (and vice versa).

1.1.10 Unemployment Rate

1.1.10.1 Definition and Data

The scoreboard indicator is the 3-year backward moving average of the unemployment rate, based on Labor Force Survey from Eurostat⁴⁷, with an indicative threshold of 10%.

Given the focus on the adjustment capacity of the economy and the ability of labor markets to reallocate labor resources, the average over the last 3 years is preferred to yearly figures which are strongly influenced by short term volatility. In this sense, the selected indicator can be seen as a proxy of the structural unemployment rate, which is, however, an unobservable variable and the estimates of which are subject to numerous caveats. Similarly, the indicator considers levels of unemployment rather than changes, as increases/drops in unemployment tend to be highly correlated with GDP growth.

1.1.10.2 Indicative Threshold

The statistical approach delivers an indicative upper threshold of 10% based on the upper quartile of the historical distribution. Due to the focus on adjustment in labor markets and not on cyclical fluctuations, only an upper threshold was considered in the scoreboard.

1.1.10.3 Complementary Indicators

This indicator should be read in conjunction with forward-looking scoreboard indicators, as its purpose is not to make unemployment as such an objective for MIP surveillance. It helps to better understand the potential severity of macroeconomic imbalances in terms of their likely persistence and the capacity of the economy to adjust.

1.1.11 Financial Sector Liabilities

1.1.11.1 Definition and Data

The scoreboard indicator is the growth rate of total financial liabilities of financial corporations, non-consolidated data, with an indicative threshold of 16.5%.⁴⁸

This indicator has the advantage of being the simplest to grasp and thus easy to communicate. The indicator is not covered explicitly by existing financial regulation and therefore there is no risk of overlap. Given that it does not require instrument disaggregation which is more prone to disentangling difficulties and reclassifications, the indicator provides also a fairly reliable basis for comparison. Moreover, it does not discriminate against different funding specificities of Member States. And last but not least, it is not specific to the business model of a specific subsector. In principle, the size of the financial sector can be measured on the asset or the liability side of financial sectors’ balance sheet. The correlation between financial assets and liabilities’ growth rates is very high (above 0.9), and thus it does not make a large difference whether financial assets or liabilities are considered.

Turning to data availability, restrictions imply that the indicator is based on nonconsolidated data (not all Member States report consolidated data) and on the annual financial accounts and balance sheets (AFA) collected by Eurostat.

1.1.11.2 Indicative Threshold

The indicative threshold of growth in financial sector liabilities is 16.5%, as derived from the upper quartile of its historical distribution. Annual data for the period 1995–2007 are used to establish the value of the threshold.

1.1.11.3 Complementary Indicators

Existent scoreboard indicators, such as credit transactions and housing price developments, already provide information on the financial sector’s efficiency of allocating resources and potential imbalances. For example, strong credit growth coupled with excessive increases of housing prices indicate a possible misallocation of credit and the build-up of an asset bubble. However, indicators capturing the change in size of financial sectors’ balance sheets, leverage indicators or soundness indicators also provide meaningful information.

Signals given by the growth rate of financial liabilities would be interpreted in the economic reading in line with developments in the real sector. For instance, if the headline indicator credit flow in the non-financial private sector (non-financial corporations and households) develops at a slower pace, one should look at what else drives growth in financial sector’s liabilities. Is this growth backed by a balanced development across liabilities’ categories, in other words, is equity growing at a similar pace with debt or is the maturity structure biased against short-term liabilities?

The debt-to-equity ratio is therefore a natural complement to be included within the list of auxiliary indicators, as the ratio indicating the relative proportion of shareholders’ equity and debt used to finance assets. The ratio provides information on the excessive leverage building up within the financial sector which can have an amplifying impact on the economic cycle (Kollmann and Zeugner 2012). A highly leveraged financial system may amplify unfavorable economic developments, like a recession, or doubts on the solvency of the sovereign. Excessive leverage by banks is widely believed to have contributed to the global financial crisis. While leverage is key for growth, excessive leverage carries the threat of the amplification effect of the volatility of returns: since the absolute increase in value of returns is accelerated when leverage is employed, so are the losses. As long as the money is optimally allocated in economic terms, high leverage levels do not necessarily imply high risk. It is only when the credibility of the borrower or the return of the underlying instrument is more uncertain that high leverage will increase the intermediary’s risk profile (European Banking Federation 2010).

	External imbalances and competitiveness					Internal imbalances
Indicator	3-year average of current account balance as a % of GDP	Net international investment position as a % of GDP	% change (3 years) of real effective exchange rate, HICP deflators relative to 41 industrial countriesa	% change (5 years) in export market shares	% change (3 years) in nominal unit labor cost b	y-o-y % change in deflated house prices c	Private sector credit flow (consolidated) as %of GDPd,e	Unemployment rate – 3-year average	Private sector debt (consolidated) as %of GDPd,e	General government sector debt as %of GDP	y-o-y % change, total financial sector liabilities, non-consolidated
Data source	EUROSTAT (BoP statistics)	EUROSTAT ( BoP statistics)	DG ECFIN (Price and Cost competitiveness).	EUROSTAT ( BoP statistics)	EUROSTAT (Nat. Accounts)	EUROSTAT	EUROSTAT (Nat. Accounts)	EUROSTAT (Labor Force Survey)	EUROSTAT (Nat. Accounts)	EUROSTAT (EDP).	EUROSTAT (Nat. Accounts)
Indicative thresholds	−4/+6% Lower quartile (also used for upper threshold)	−35% Lower quartile	+/−5% for €A +/−11% non-€A Lower, Upper Quartiles of EA −/+ s.d. of EA	−6% Lower quartile	+9% €A +12% non-€A Upper Quartile €A +3%	+6% Upper quartile	+15% Upper Quartile	+10%	133% Upper Quartile	+60%	16.5%
Additional indicators for economic reading	Net lending/ borrowing vis-à-vis ROW as % of GDP	Net External Debt as % GDP; nward FDI flows and stocks as % of GDP	Real effective exchange rate vis-à-vis rest of the euro area	Relative export market shares relative to advanced economies; Labor productivity; Trend TFP growth	Nominal ulc (changes over 1, 5, 10 years); Effective ulc relative to the rest of EA	Real house price changes (over 3 years); Nominal house price index Value-added in residential construction	Change in private debt	Participation rate, long-term and youth unemployment poverty indicators	Private sector debt based on non-consolidated data		Debt over equity ratio

1. ^aNotes: For EU trading partners HICP is used while for non-EU trading partners, the deflator is based on a CPI
2. ^bIndex providing ratio of nominal compensation per employee to real GDP per person employed
3. ^cChanges in house prices relative to the consumption deflator
4. ^dPrivate sector is defined as non-financial corporations; households and non-profit institutions serving households
5. ^eSum of loans, and securities other than shares

1.2 Annex A.2.2 Scoreboard indicators and their indicative thresholds

1.3 Annex A.2.3 MIP Scoreboard in AMR-2015 (Published End 2014 with Values Up to 2013)

Table 3

1.4 Annex A.2.4 Auxiliary Indicators in the MIP Scoreboard in AMR-2015 (Published End 2014 with Values Up to 2013)

Table 4

Table 5

Table 6