Although the source of the Global Crisis of 2008 was the US, Europe has suffered the most in terms of the depth of the recession and, more importantly, in terms of a slow recovery. Financial markets and the banking sectors in several European countries are still struggling to recover from the crisis, with many instances of unresolved problems involving bad loans and fragility of banks. The chapters in a new CEPR Press eBook, resulting from the first conference of the European Central Banking Network (ECBN), held in Ljubljana in October 2015, focus on the role played by the financial sector in the allocation of resources across different firms and sectors of the economy (Banerjee and Coricelli 2017).
The papers presented at the conference discussed whether misallocation is magnified during credit booms and whether misallocation is reduced during the deleveraging process following a financial crisis. The conference provided a unique perspective, covering a broad sample of countries characterised by different levels of development of financial markets, different magnitudes of macroeconomic imbalances, and different policy responses.
In this column, we provide a broad overview of some main stylised facts for a large sample of European countries.1
We focus on the relationship between credit booms and busts and the potential misallocation of resources at the micro level. Some of the hardest hit countries in Europe experienced a pre-crisis credit boom followed by deleveraging and, in some cases, a creditless recovery. The key questions are:
- How did the credit boom affect the efficiency of the system in terms of allocation?
- How is the deleveraging affecting the efficiency of allocation?
When financial markets are imperfect, the allocation of resources may be inefficient. Furthermore, with financial imperfections, deleveraging may not lead to a more efficient allocation, even when the pre-crisis boom was highly inefficient. For instance, reliance on collateral implies that the allocation of credit follows a collateral criterion rather than efficiency/productivity of the borrowing firm. The possible inefficiencies of the credit boom preceding the Great Recession have often been associated with macro imbalances and sectoral imbalances. In Southern Europe and some Central and Eastern European countries, macro imbalances went hand-in-hand with sectoral imbalances. The accumulation of huge current account deficits was associated with a boom in non-tradable sectors, especially real estate and construction, accompanied by shrinking shares of manufacturing output and employment. Less attention has been given to more micro measures of misallocation, which are however crucial for assessing the costs of the crisis and the prospects for recovery. This is crucial to determine whether the recession produced a shift to a lower potential output.
We use the Amadeus firm-level database to give a broad picture of the behaviour of misallocation of resources before and after the Great Recession. We look at both within-sector allocation and across-sectors allocation. Following Hsieh and Klenow (2009), as is done in many of the chapters in this book, we define misallocation as arising from two sources: different productivities across firms within the same sector, and inefficiency in the allocation of inputs across firms.
Because of frictions and policy distortions (taxes, financial frictions), a significant fraction of productive resources (inputs) are employed in low-productivity firms, instead of being employed in high-productivity firms. Misallocation of resources is thus detected by observing a distribution of firms within sectors with a large dispersion and fat tails. In the distribution chart, a fat left tail signals a large weight of low-productivity firms. The main question addressed in this book is whether inefficiencies in financial markets and financial cycles have an impact on the degree of misallocation of resources? As noted by Restuccia and Rogerson (2008), "[f]avoured establishments demand more capital and become larger than in the absence of the distortion”.
Using the Amadeus database, we also compute misallocation separately for various clusters: industry, country, time period and macro-regions.2
Misallocation can be visually summarised by looking at the distribution of total factor productivity in firms within finely defined sectors. The case of no misallocation would correspond to a distribution collapsing to a line at zero. Looking at macro regions, we find higher misallocation in Eastern Europe. In terms of sectors, there is higher misallocation in services, possibly reflecting a lower degree of competition here than in manufacturing. The dynamics over time indicate little change before the crisis, some change during the crisis and also some change after it.
Measuring misallocation at the country level, we find that misallocation plays a crucial quantitative role in explaining productivity differences across countries. Figure 1 reports as an example the distributions of total factor productivity (TFP) for Germany, Italy and Ukraine.3 Italy, for instance, could increase its industrial TFP by 7.5% if its allocative efficiency were aligned to that of Germany.
Figure 1 Distribution of TFP by country
Note: TFPR indicates revenue total factor productivity, which is real total factor productivity multiplied by output prices.
Addressing the role of the Global Crisis, Figure 2 summarises the distributions for the period before 2008 and after it. We note that the change in misallocation during 2004-2007 corresponds to a TFP gain of 0.35%, while during the crisis period of 2007-2011 the increase in misallocation corresponds to a reduction in TFP of 0.9%.
Figure 2 Misallocation over time, whole sample (%)
We then distinguish countries in terms of the dynamics of credit before the crisis. We use the methodology from Gourinchas et al. (2001) to identify credit boom episodes by looking at deviations from trends. For the two macro-regions Figure 3 reports the difference in the MIS index in credit boom countries versus no-credit boom countries. Results indicate that credit booms are associated with higher misallocation in Western Europe but not in Eastern Europe, while the crisis led to a closing of the gap in Western Europe and a widening in Eastern Europe.
Figure 3 Misallocation: Differences between credit booms vs non-credit booms
We analyse more deeply the role of credit booms by running a regression that allows us to control for ‘excessive’ debt exposure before the crisis. We then interact this variable with the credit boom dummy, which allows us to better identify the effects of the credit boom, by controlling for the different excessive debt accumulation by sectors. Excess debt is computed as the ratio of debt to capital in a given sector/country relative to the average for the whole sample. Therefore, we try to capture the fact that credit booms disproportionally affected misallocation of resources in sectors that displayed the largest debt exposure, relative to the European mean.
Results indicate that credit booms induced a significant increase in misallocation in Western Europe but not in Eastern Europe. One possible explanation is that in Eastern Europe, firms might still be far from an optimal level of indebtedness and thus there is still room for an increase in debt-to-capital ratios that reflects an equilibrium phenomenon rather than an inefficient ‘excess’. Note that this does not contradict findings on misallocation across sectors, with an excessive accumulation of debt in non-tradable sectors in Eastern Europe.
In summary, credit booms seem to be associated with higher misallocation, even within sectors. This is particularly true for Western Europe, but less so for Eastern Europe.
Overall, the crisis has not brought any visible improvement in the allocation of resources. Therefore, there is no evidence that deleveraging has had, at least initially (up to 2011), any ‘cleansing’ effects. Improving the functioning of credit markets and their ability to improve the allocation of resources is crucial to lift Europe out of the phase of low growth that has followed the Global Crisis.
Banerjee, B and F Coricelli (2017), Crisis, Credit and Resource Misallocation: Evidence from Europe during the Great Recession, CEPR Press.
Coricelli, F. and M. Frigerio (2016), “Credit boom and bust in Europe: Dynamic of misallocation during the Great Recession,” Paris School of Economics, mimeo.
Gourinchas P.O, R. Valdes and O. Landerretche (2001), “Lending booms: Latin America and the world”, Economia Journal of LACEA 1(2): 47-100.
Hsieh, C. and P. Klenow (2009), “Misallocation and Manufacturing TFP in China and India,” Quarterly Journal of Economics 124: 1403-1448.
Restuccia, D. and R. Rogerson (2008), “Policy Distortions and Aggregate Productivity with Heterogeneous Plants”, Review of Economic Dynamics 11: 707-720.
1 The empirical analysis summarised in this column draws from Coricelli and Frigerio (2016).
2 Western Europe; Central Eastern Europe plus Turkey and Cyprus, which for simplicity we call “Eastern Europe”. Industry disaggregation is at the 4-digit level.
3 Our estimates for advanced EU countries are of the same order of magnitude of those estimated by Hsieh and Klenow (2009) for the US.