The Essential Characteristic of a Central Bank – Independence 69

Central Bank Journal of Law and Finance

Year II, no. 1, 2015

The Central Bank Journal of Law and Finance is coordinated by the National Bank of Romania – Legal Department, Director Alexandru-Nicolae Păunescu, PhD 1. SCIENTIFIC BOARD Mugur ISĂRESCU Academician, Romanian Academy, Corresponding member of Real Academia de Ciencias Económicas y Financieras, Barcelona, Spain Moisă ALTĂR Professor, DoFIN, Academy of Economic Studies, Bucharest, Romania, Romanian-American University, Bucharest, Romania Lucian BERCEA Professor, West University of Timişoara, Romania Radu CATANĂ Professor, Babeş Bolyai University, Cluj, Romania James Ming CHEN Professor, Justin Smith Morrill Chair in Law, Michigan State University, USA Ionuţ DUMITRU Professor, DoFIN, Academy of Economic Studies, Bucharest, Romania Ovidiu FOLCUŢ Professor, Romanian-American University, Bucharest, Romania M. Peter van der HOEK Professor, Academy of Economic Studies, Bucharest, Romania, Erasmus University (EM), Rotterdam, Netherlands Iftekhar HASAN Professor, E. Gerald Corrigan Chair in International Business and Finance, Fordham University’s Schools of Business, New York, USA Dumitru MIRON Professor, Academy of Economic Studies, Bucharest, Romania Ion STANCU Professor, Academy of Economic Studies, Bucharest, Romania Brânduşa ŞTEFĂNESCU Professor Emeritus Paul WACHTEL Professor, Stern School of Business, New York University, USA

2. EDITORIAL BOARD Tudor CIUMARA PhD, Researcher, Centre of Financial and Monetary Research, Bucharest, Romania Adrian - Ionuţ PhD, Associate Professor, Academy of Economic Studies, CODIRLAŞU Bucharest, Romania Bogdan-Octavian PhD, Senior Lecturer, Academy of Economic Studies, COZMÂNCĂ Bucharest, Romania Adina CRISTE PhD, Researcher, Centre of Financial and Monetary Research, Bucharest, Romania Constantin MARIN PhD, Researcher, Centre of Financial and Monetary Research, Bucharest, Romania Beatrice POPESCU Adviser to the Deputy Governor of the National Bank of Romania Radu RIZOIU PhD, Lecturer, Faculty of Law, University of Bucharest, Romania

iii

Cover by Romeo Cȋrjan, PhD

DISCLAIMER

The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the National Bank of Romania, nor do they engage it in any way.

ISSN 2392 – 9723 ISSN-L 2392 – 9723

CONTENTS

James Ming Chen The Promise and the Peril of Parametric Value at-Risk (VaR) Analysis 1

Andreas Guericke, Jochen Sprung Designing Macroprudential Mandates and Institutional Frameworks: The German Approach 43

Lucian Croitoru Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan) 57

Ianfred Silberstein The Essential Characteristic of a Central Bank – Independence 69

Mirela Iovu The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency 89

Irina Mihai, Alina Tarţa The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania 107

Dan Pălăngean The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective 133

AUTHOR GUIDELINES for CENTRAL BANK JOURNAL OF LAW AND FINANCE

The future contributors of the Central Bank Journal of Law and Finance, published under the auspices of the National Bank of Romania, are requested to consider the guidelines below.

DRAFTING GUIDELINES Manuscript Formatting. Prepare the manuscript in English, using Georgia 10 pt. for the body text and 9 pt. for footnotes and legends, 1.2 line spacing, A4 paper size and submit it as doc, docx or rtf. Do not use bold, underline or ALL CAPS. Use italics sparingly, for emphasis only. Do not use tabs or indents.

Manuscript Structure. Structure the manuscript as follows: the manuscript title in capital letters, an abstract that should not exceed 100 words, keywords, JEL classification, main text, acknowledgements, appendices, references. Divide your manuscript into clearly defined and numbered sections.

Tables, Graphics and Schemes. Provide numbered captions for all tables and figures.

Citations. Use endnotes to cite sources. Citations should follow the format: author’s last name, author’s first name, full title of the cited paper, publisher, place of publication, year, cited page(s).

References. Submit each manuscript with a full list of references. The list should be numbered, delimited by semicolons and each reference should follow the format: author’s last name, author’s first name, full title (do not use abbreviations), publisher, place of publication, year, cited page(s).

SUBMISSION Submit your manuscript to the e-mail address: [email protected].

The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

James Ming Chen*

Abstract

Leptokurtosis, or the risk lurking in “fat tails,” poses the deepest epistemic threat to economic forecasting. Parametric value-at-risk (VaR) models are vulnerable to kurtosis in excess of the levels associated with a Gaussian distribution. Careful application of Student’s t-distribution can enhance the statistical robustness of VaR forecasts. For degrees of freedom greater than 4, Student’s t-distribution can emulate any level of excess kurtosis. Observed levels of kurtosis guide the use of Student’s t-distribution to measure VaR. Parametric VaR according to the number of degrees of freedom implied by historical levels of kurtosis also yields the corresponding value of expected shortfall. This parametric technique not only exploits the elicitability of VaR from historical data, but also informs the computation of expected shortfall as a coherent risk measure.

Keywords kurtosis, value-at-risk, VaR, expected shortfall, Basel accords, Student’s t-distribution

JEL Classification: C58, G17, G21

* Justin Smith Morrill Chair in Law, Michigan State University. I presented this paper on June 24, 2015, to the Faculty of Economics of the University of Zagreb (Ekonomski Fakultet Sveučilišta u Zagrebu). I have benefited from scholarly interactions with Anna Agrapetidou, Moisa Altar, Chris Brummer, Adam Candeub, Robert Dubois, John F. Duffy, Santanu Ganguli, Tomislav Gelo, Periklis Gogas, Jagoda Kaszowska, Daniel Martin Katz, Yuri Katz, Imre Kondor, Othmar Lehner, Milivoj Marković, José María Montero Lorenzo, Vivian Okere, Merav Ozair, Carol Royal, Jeffrey A. Sexton, Jurica Šimurina, Nika Sokol Šimurina, Robert Sonora, Benjamin Walther, Gal Zahavi, and Johanna F. Ziegel. Christian Diego Alcocer Argüello, Emily Strickler, and Michael Joseph Yassay provided very capable research assistance. Special thanks to Heather Elaine Worland Chen.

Central Bank Journal of Law and Finance, No. 1/2015 1 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

T.S. Eliot, the celebrated Anglo-American poet, called history “a pattern/ Of timeless moments.”1 In projecting forward rather than backward over time, economic forecasting cannot escape moments of a different sort: the mathematical moments of the distribution of financial returns. Of the four moments of greatest interest to financial institutions and their regulators2 — mean, variance, skewness, and kurtosis — it is the fourth moment, kurtosis, that should pose the deepest epistemic concern. Kurtosis eludes detection where it counts most — in its fat tails. Our expectations and perceptions may underestimate the most extreme risks by a significant margin. The overarching goal in financial responses to leptokurtosis and “fat tails” is the accurate forecasting of extreme events. Simple accuracy in description, if attainable and attained, would be a fantastic accomplishment. This article will describe one approach for managing kurtosis in financial returns. Despite its flaws, conventional value-at-risk (VaR) analysis remains a vital tool in financial risk management. VaR and related risk management techniques have been embraced by the Basel Committee on Banking Supervision and codified in the accords known as Basel II, 2.5, and III. But parametric VaR models are extremely vulnerable to kurtosis in excess of the levels associated with a normal, Gaussian distribution. Using a systematically leptokurtic distribution to forecast extreme losses provides a workable response to the model risk that haunts VaR analysis. This article provides step-by-step guidance on the use of Student’s t-distribution as an easily implemented model for enhancing the robustness of VaR forecasts. Although the distribution and quantile functions of Student’s t-distribution lack simple, closed-form representations,3 the distribution’s even-numbered moments are extremely tractable. For degrees of freedom greater than 4, Student’s t-distribution can emulate any level of kurtosis exceeding that of a Gaussian distribution. The elicitability of VaR from historical data invites the use of observed kurtosis as a credible theoretical basis for enhancing the robustness of VaR measurements. Part I of this article provides background on kurtosis. Mindful of the ongoing significance of conventional VaR in financial risk management, Part II describes parametric VaR on a Gaussian model. Part III implements an alternative approach to parametric VaR according to Student’s t-distribution. The pivotal question is determining the number of degrees of freedom. As an admonition to heed the shape as well as the scale of a statistical distribution, part IV will compare Student’s t-distribution to the logistic distribution, whose excess kurtosis is even more easily calculated. In concert, parts III and IV illustrate techniques that respond to Basel 2.5’s prescription of stressed VaR, or value-at-risk analysis under conditions of market stress. Part V extends these exercises in stressed VaR to expected shortfall, the risk measure endorsed by Basel III. Part VI concludes with further cautionary observations on the limitations of parametric VaR analysis.

2 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

1. LEPTOKURTOSIS, FAT TAILS, AND SUPER-GAUSSIAN DISTRIBUTIONS

Excess kurtosis is formally “defined as the fourth cumulant divided by the square of the second cumulant, which is equal to the fourth moment around the mean divided by the square of the variance of the probability distribution minus 3”:4

4 4  2  2  4  3 2  One way to tame this formidable mathematical definition is to render it slightly more manageable with Greek neologisms. Excess kurtosis describes the shape of a probability  distribution. Mesokurtic distributions have precisely the kurtosis of the standard normal distribution. Their excess kurtosis is therefore 0. Platykurtic distributions, such as the uniform and Bernoulli distributions, have less kurtosis relative to the Gaussian baseline. The class of distributions of greatest interest in financial risk management is the super-Gaussian class of leptokurtic distributions. These distributions are characterized by a thin “peak” surrounding the mean and fat tails. “Examples of leptokurtic distributions include the Cauchy distribution, Student’s t-distribution, Rayleigh distribution, Laplace distribution, exponential distribution, Poisson distribution and the logistic distribution.”5 Even this nomenclature has the potential to confuse. Λεπτός, the root of leptokurtic and leptokurtosis, means “slender,” the very opposite of the trait that matters most in financial risk management: the fatness of a leptokurtic distribution’s tails. An alternative taxonomy, mercifully rendered in plain language, emphasizes the tails.6 A thin-tailed distribution such as the uniform distribution has a finite upper limit. A medium-tailed distribution such as the Gaussian distribution has exponentially declining tails. A fat-tailed distribution such as Student’s t-distribution or the logistic distribution has power-law tails. Leptokurtosis matters because neither intuition nor conventional statistics can prepare financial institutions or their regulators to evaluate low-probability, high-impact events. Underestimating the likelihood and impact of loss poses especially treacherous risks in finance.7 Leptokurtic “probability distributions are inherently difficult to estimate.”8 Indeed, when “events are rare,” the sheer lag time between iterations often makes it “impossible to estimate just how quickly the tail tapers off.”9 Because “fat tails bring with them an epistemic problem,”10 it is wholly unsurprising that we lack “a commonly accepted economic framework” for responding to them.11 At an extreme, the problem threatens to remove the entire class of problems best modeled by fat-tailed distributions from the category of risk, where probability is quantifiable, and to sweep such problems into the distinct category of uncertainty, where probability is unquantifiable.12 Fat tails define a zone of “[t]rue

Central Bank Journal of Law and Finance, No. 1/2015 3 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis uncertainty,” where risks “are not well understood, where the range of outcomes is potentially very large, and where probabilities cannot be assigned with confidence.”13

2. PARAMETRIC VALUE-AT-RISK (VAR) ANALYSIS

2.1. The Vulnerability of VaR Analysis to Model Risk

Leptokurtosis poses an especially keen threat to the economically informed evaluation of market risk in the trading books of systemically important financial institutions. Value-at- risk analysis arguably remains the most important tool for evaluating market risk within the global financial system.14 Even though the Basel Committee on Banking Regulation has contemplated alternative measures of market risk,15 the Basel II accord did endorse a version of VaR analysis as its preferred tool for risk assessment.16 Many systems of risk management, whether through regulation or through voluntarily adopted best practices, have embraced some form of VaR analysis.17 Authorities around the world have endorsed VaR, either as a binding regulatory standard or as a voluntary best practice.18 Even absent legal compulsion, private firms routinely use VaR as an internal risk management tool, often directing traders to reduce exposure below the level prescribed by those firms’ own VaR limits.19 In turn, regulators have “assimilated” the financial industry’s “own internal risk models,” including proprietary VaR algorithms, into supervisory capital requirements and other legal standards.20 Such practices are undoubtedly facilitated by the release of the original RiskMetrics specification of VaR into the public domain and the subsequent widespread circulation of that methodology among academics and financial professionals.21 Like any other econometric technique, VaR is subject to the risk that its underlying model miscalculates the level and nature of actual risk.22 Despite serious (and empirically warranted) doubts about the applicability of the Gaussian distribution to financial returns,23 the simplest parametric implementations of VaR assume normally distributed returns.24 To compound the method’s problems, VaR often relies on strictly historical data.25 VaR falls victim to its own elegance. The “simplicity of VaR measures,” part of this methodology’s appeal to quantitative analysts and to regulators, “is in large part obtained with assumptions not supported by empirical evidence.”26 Of these assumptions, the “most important (and most problematic) … is that returns are normally distributed.”27 This assumption, shared by many other applications of mathematical finance,28 exposes VaR and its practitioners to levels of unforeseen tail risk. Whatever else they do, stock market returns do not follow the normal distribution.29 Therefore, reliance on a normal, Gaussian distribution systematically understates the market risk borne by any portfolio.

4 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

Departures from expected value, especially in the tails of a distribution, put a premium on statistical robustness — the resistance of a statistical model to outliers or other deviations from the model’s underlying assumptions.30 One easily implemented way of heightening the robustness of VaR analysis in anticipation of unobservable and therefore unpredictable tail risks is to recalibrate parametric VaR according to a more leptokurtic distribution. To fully appreciate the value of such a technique, however, we should first establish the methodological baseline of conducting parametric VaR according to a Gaussian model. I therefore turn to that task.

2.2. Parametric VaR According to a Gaussian Distribution

Let us begin by performing a simplified version of parametric VaR analysis. Suppose that an investor stakes $1 million on an index fund tracking the Standard & Poor’s 500.31 She asks her financial advisor, “If capital markets go down to an extent witnessed only once in a hundred trading days, what can I lose by tomorrow’s market close?” Adjusted to the appropriate scale, this problem represents virtually every question of financial market risk. VaR analysis has supplied financial actors and their regulators with a facile tool for calculating the portion of a portfolio that may decline over some interval of time. To answer our investor’s question, an advisor using conventional parametric VaR may assume a mean daily return of 0, with a standard deviation over that interval of 100 basis points (equal to 1 percent). On those assumptions, that advisor will report a one-day value of

VaR0.01 as $23,264 for a $1 million portfolio. VaR0.01 = $23,264 is a fancy, technocratic way of telling this investor that she faces a 1 percent chance of losing at least $23,264 on her S&P 500 index fund on any given trading day. Equivalently, the advisor could tell the investor client that her portfolio has a 99 percent chance tomorrow (after a single trading day) of being worth at least $976,736 ($1,000,000 – $23,264). In formal terms, VaR for a certain risk or confidence level is the quantile that satisfies the confidence level, , in the following equation:32

VaR    f (x) dx 

In the case of the investor with a $1 million portfolio invested in an S&P 500 index fund,  = 1 – 0.01, or 0.99. f(x) refers to the probability density function — in this case, of the  distribution of returns on the investor’s S&P 500 fund. VaR may also be defined as the greatest lower bound (infimum) on the cumulative distribution function F of any financial position Y, expressed as a real-valued, random variable:33

VaR (Y)  inf{x  | FY (x) }

 Central Bank Journal of Law and Finance, No. 1/2015 5 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

Parametric VaR analysis requires the computation of statistical quantiles.34 The quantile function of a distribution is the inverse of its cumulative distribution function, which in turn is the definite integral of the probability density function. For normally distributed returns, the quantile function is designated by the inverse of the capital phi symbol that designates the cumulative distribution function of the Gaussian distribution: –1(p). The quantile function of the standard normal distribution, also known as the probit function,35 is expressed as a transformation of the inverse error function:36

1 1 zp   (p)  2 erf (2p 1)

Conventional VaR notation designates the quantile function as zp. Devising four alternate ways for referring to the same mathematical concept — quantile function, inverse cumulative  distribution function, probit, and zp — may intuitively boost the understanding of the quantitative mechanics at work. Formally, “[t]he quantile zp represents such a value that a

standard normal random variable X has the probability of exactly p to fall inside the (–∞, zp] interval.”37 In effect, we are asking what standard score, or z,38 corresponds to the value of the cumulative distribution function representing a certain percentage of the total under the curve that defines the probability density function of the returns on an investment. We have now assembled the tools needed to complete our simple VaR analysis. Recall that we have assumed our investor has staked $1 million on an S&P 500 index fund, where mean daily return (μ) is 0 and the standard deviation of that mean return (σ) is 100 bps (0.01). The variable VaRp expresses the value at risk given a particular probability of a loss as the product of −zp, standard deviation σ, and the total value of the portfolio (v):39

VaRp  zp    v

The negative sign before −zp reports value at risk as a positive sum. For σ = 100 bps and v = $1,000,000:  VaR0.01  z0.01   v

VaR0.01  z0.01 100 bps  $1,000,000

So far we have omitted any consideration of time. As long as returns are independent and identically distributed (a crucial assumption of any distribution obeying the central limit  theorem),40 “variances are additive over time, which implies that volatility grows with the square root of time.”41 To account for variance over time, we typically multiply VaR by the square root of time:42

VaRp, t  zp    v  t

 6 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

Adopting the simplifying assumption of a single trading day, √t = 1, enables us to forgo the potential complications that attend the consideration of time in VaR analysis.43

All that stands between us and a complete calculation of VaR0.01 is the value of z0.01. That value in turn requires the application of the quantile function:

1 1 z0.01   (0.01)  2 erf (2 0.01 1)  2.3264

Inserting this value of z.01 into the formula for VaR0.01 yields the conclusion that VaR0.01 for this asset, over a trading interval of a single day, is approximately $23,264. The following  table expresses cumulative probabilities for the foregoing exercise in parametric VaR analysis at commonly used intervals:44

Table 1

P 0.1% 0.5% 1.0% 2.5% 5.0% 10%

Zp –3.0902 –2.5758 –2.3264 –1.9597 –1.6449 –1.2816

VaRp $30,902 $25,758 $23,264 $19,597 $16,449 $12,816

3. PARAMETRIC VAR ACCORDING TO STUDENT’S T-DISTRIBUTION

3.1. Choosing Among Non-Gaussian Distributions

Parametric VaR “generalizes to other distributions as long as all the uncertainty is contained in σ.”45 If we are concerned that reliance on the Gaussian distribution systematically and inappropriately underestimates tail risk, we could substitute any “distribution [with] fatter tails than the normal.”46 Phillipe Jorion recommends the use of Student’s t-distribution with six degrees of freedom.47 I shall do my best to provide a theoretical justification for conducting parametric VaR according to the family of Student’s t-distributions, and to suggest a basis for securing empirical support for such a choice. There is an inevitable element of arbitrariness in the choice of a statistical distribution as the basis for a model of financial risk. Any distribution of returns could provide a credible economic basis for parametric VaR. The use of elliptical distributions enables analysts to characterize all portfolios entirely by location and scale,48 so that a single synthetic portfolio of a particular location and scale can model any actual portfolio predicted to have those parameters.49 The most basic statistical parameters associated with location and scale are

Central Bank Journal of Law and Finance, No. 1/2015 7 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis those drawn, respectively, from the first and second mathematical moments of any distribution: the mean and the variance, whose positive square root is the standard deviation. Half a century ago, the financial academy might well have addressed leptokurtosis by resort to the family of stable distributions.50 “A Lévy stable distribution can have skewness and fat tails and obeys scaling properties.”51 In 1963, both Benoit Mandelbrot52 and Eugene Fama53 hypothesized that stock prices follow a stable distribution. Fama later asserted that “[t]he presence, in general of leptokurtosis in the empirical distributions seems indisputable.”54 At least where financial returns stubbornly refuse to exhibit finite variance,55 the appropriate statistical model for asset prices may indeed be the stable distribution.56 These traits led Mandelbrot to posit that financial markets have infinite variance, perhaps even infinite expected value.57 Despite their prominence in the foundational literature of mathematical finance, stable distributions “have fallen out of favor … because they make financial modeling so difficult.”58 The trouble is that for any value of shape parameter α < 2, the stable distribution has infinite variance.59 The “infinite variance” of most stable distributions not only “violates empirical observations and logic that dictates finite variance,” but also “significantly complicates the task of risk estimation” and limits the practical application of the stable distribution.60 When α = 2, we encounter a special case of the stable distribution called the normal distribution.61 For most other stable distributions, however, the usual statistical parameters, from mean and standard deviation to skewness and excess kurtosis, are mathematically undefined.62 The Cauchy distribution, another special case of the stable distribution, has scale parameters α = 1 and β = 0.63 The Cauchy distribution, with an undefined mean, infinite variance, no finite moments of order greater than zero, and no moment-generating functions, is applied mathematics’ “canonical example of a ‘pathological’ distribution.”64 In practice, financial forecasting is more tractable than Mandelbrot and Fama suggested. “[S]ample estimates of …variance and higher moments” converge rather than “increase as sample size increases.”65 Moreover, evidence of abnormality “is much weaker for long- horizon returns than for short-horizon returns.”66 Both of these traits suggest that the law of large numbers and the central limit theorem remain viable: a sufficiently large number of independent and identically distributed random variables, each with finite mean and variance, will be approximately normally distributed.67 Contemporary financial analysis therefore prefers distributions with “finite second moments of returns, and often finite higher moments as well.”68 It nevertheless bears remembering that sheer sample size cannot overcome the problem of leptokurtosis on its own. After all, the economic maxim, “In the long run we are all dead,”69 is drawn from an immutable biological reality. In the face of firm limits on the ability to gather a sufficient large number of iterations, one obvious response is to “model [financial]

8 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen returns as drawn from … fat-tailed distribution[s] with finite higher moments, such as the t distribution.”70

3.2. An Introduction to Student’s t-Distribution

Student’s t-distribution has very attractive properties that make it an ideal model for parametric VaR.71 Even though financial models based on Student’s t-distribution “are unstable” in the sense that “their shapes change at different time horizons and that distributions at different time horizons do not obey scaling relations,” these models do “possess finite variance and fat tails.”72 The entire family of multivariate Student’s t- distributions belongs to the broader class of elliptical distributions.73 The lone parameter of Student’s t-distribution, , or the number of degrees of freedom, enables easy adjustment of the tails to reflect a wide range of leptokurtic distributions:

Figure 1: A plot of the Probability Density Function for Several Members of the Student’s t Family of Distributions

Source: https://en.wikipedia.org/wiki/Student%27s_t-distribution#/media/File:Student_t_pdf.svg

As we will see, adjusting the  parameter generates an entire family of Student’s t- distributions, spanning everything from the normal, Gaussian distribution to the pathological Cauchy distribution. My primary purpose here is not to conduct a clinic on ways to enhance the robustness of financial models relative to that of the Gaussian distribution.74 Rather, I wish to demonstrate useful modifications of conventional parametric value-at-risk analysis. For these purposes, Student’s t-distribution offers two principal advantages over its Gaussian counterpart. The first advantage is one of similarity. Both distributions are symmetrical and centered on a

Central Bank Journal of Law and Finance, No. 1/2015 9 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

mean of 0. Formally, skewness for both distributions is 0. Although real-world risk assessment often demands strict reliance on empirical evidence of risk and return and nonparametric methodology, modeling still matters. Of course, there are statistical distributions that exhibit both skewness and leptokurtosis relative to the Gaussian baseline. It is possible to introduce skewness into any continuous, unimodal distribution that is symmetric with respect to 0.75 It is also possible to conduct parametric VaR analysis using other distributions, such as a three-parameter lognormal distribution76 or a log-logistic/Fisk distribution,77 which are both asymmetrical and fat-tailed. Indeed, it is possible to organize entire families of statistical distributions according to mean, variance, and skewness78 — or mean, variance, skewness, and kurtosis79 — and to select among these distributions according to algorithms that report the best fit for a given data set.80 For the purpose of outlining a tractable approach to parametric VaR, we are better served, at least initially, to isolate issues of excess kurtosis, wholly apart from skewness, and to evaluate models solely on the basis of differences in kurtosis. The second, and primary, advantage of Student’s t-distribution over the Gaussian distribution is a higher level of kurtosis. The tails of Student’s t-distribution are fatter. In the illustration above, Student’s t-distribution opens up perceptibly, relative to the Gaussian distribution, especially at values for |z| exceeding 2σ. The “ability to account for the observed kurtosis of empirical distributions of daily rates of return” historically drew the attention of mathematical finance to the stable distribution.81 Student’s t-distribution, however, boasts a crucial advantage over the stable distribution. Student’s t-distribution addresses elevated levels of kurtosis without “indicating a distribution whose theoretical second moment is infinite.”82 This trait makes Student’s t-distribution a popular choice for robust parametric modeling.83 The specifications of Student’s t-distribution’s probability density and cumulative distribution functions are analytically and computationally formidable. Formally, Student’s t-distribution has the following probability function:

 1 ( 1)  t 2  2 2 f (t)   1  ( 2 )  

where  represents the number of degrees of freedom and  represents the gamma function.84

 Rewriting the probability density function in terms of the beta function (β)85 will clarify the usual specification of the cumulative distribution function:

10 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

 1 1  t 2  2

f (t)  1  1  ( 2 , 2 )  

I, the regularized incomplete beta function,86 is used to specify the cumulative distribution function of Student’s t-distribution:

 t 1  1 F(t)   f (u) du 1 2 Ix(t )( 2 , 2 )  where   x(t)  t 2  The quantile or inverse cumulative distribution function of Student’s t-distribution, F–1(t), which is pivotal to VaR analysis, takes the following form:

 1 zp  F (t) 

1 1   1  1 1, 0  t  2 I2t ( 2 , 2 ) 0, t  1 2

1 1   1  1 1, 2  t 1 I2(1t )( 2 , 2 ) , t  0 , t 1

1 87 where Ix (a, b) indicates the inverse regularized beta function.

 On the other hand, extremely simple algorithms enable the calculation of variance and excess kurtosis in Student’s t-distribution. Raw moments of Student’s t-distribution exist only for   > 1 and for orders less than . All odd moments are 0. For even values of k, 0 < k < , the following function describes the central moment of order k:

k 2 k 2 j 1 E(T k )   2  j 1   2 j

A moment of order k exists only for a t-distribution whose number of degrees of freedom exceed that order:  > k. For values of  in the range, 2 <  < 4, variance alone is defined. 

Central Bank Journal of Law and Finance, No. 1/2015 11 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

Variance and excess kurtosis are both defined if and only if  > 4. In that event, variance, standard deviation (as the square roof of variance), and excess kurtosis are all remarkably simple:  2    2      2 6   2   4 The theoretical significance of the foregoing formulas is profound. At a sufficiently large number of degrees of freedom, but not so many as to merge with the normal distribution,  Student’s t-distribution provides an easily implemented model for observed kurtosis in market returns. For  > 4, Student’s t-distribution reports finite, tractable values for variance and excess kurtosis. Without conceding that “rates of return do, in fact, follow a Student model,” an early comparison of leptokurtic distributions concluded “that the Student model provides a better empirical fit” for observed returns “than the stable model.”88 It is also very helpful that extreme cases of Student’s t-distribution correspond to distributions at opposite ends of the elliptical set. For  = 1, Student’s t-distribution becomes the Cauchy distribution, the pathological stable distribution with infinite variance and kurtosis: 1 f (x)  (1 x 2 )

1 tan1(x) F(x)   2  By contrast, as   ∞, Student’s t-distribution approaches a normal distribution. When  = ∞, Student’s t-distribution becomes the normal distribution. Although modeling returns  according to Student’s t-distribution does imply that returns “converge[] to normality for large sum sizes,” actual evidence indicating convergence to normality in empirical returns ultimately “give[s] additional support” to a preference for Student’s t-distribution over stable distributions whose variance and kurtosis are undefined.89 All that remains before the application of Student’s t-distribution to parametric VaR is the determination of the  parameter. In conventional applications of Student’s t-distribution, the number of degrees of freedom is typically the size of the sample, minus 1 ( = n – 1). The

12 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

t-distribution corresponding to  degrees of freedom therefore defines the distribution about the true mean, relative to the sample mean and divided by the square root of the sample size, √n, as a normalizing factor: x   t  s/ n

where x is the sample mean and s2 is the sample variance. At a sufficiently high level of theoretical abstraction, the use for which we are conscripting  Student’s t-distribution does fit that distribution’s original purpose. Even with hundreds or  thousands of observations, the evaluation of potentially catastrophic financial events is necessarily starved for data. In this light, it makes sense to use small-sample versions of Student’s t-distribution to estimate the range within which a current point estimate of wealth or risk represents a reliable forecast of actual future wealth or risk. For our unconventional use of Student’s t-distribution, sample population n provides no guidance on the value of . We do know that  must be greater than 4, if we hope to have a defined, finite value for excess kurtosis. In setting the value of  on some basis besides raw

aesthetics, we should pay attention to the corresponding value of 2. In computing parametric VaR on the basis of Student’s t-distribution, we should choose the value of 

according to the corresponding level of excess kurtosis. For  ≤ 4, 2 = ∞. Mercifully, 2  however, variance is defined at  = 4:    2  2;   2 . Demanding that  > 4 therefore sets the lower boundary on values of  corresponding to finite variance and finite kurtosis. To be sure, using noninteger values for  is inconsistent with conventional applications of Student’s t-distribution, where the goal is to define the distribution of the true mean of the entire population, relative to the mean of a limited sample. But the probability density function of Student’s t-distribution supports all values of  > 0; its only limitation is that central moments of order k are defined only to the extent that k > . The condition that  > 4

therefore defines the lower boundary of our exercise in parametric VaR. As   4, 2  ∞.

At  = 5, we encounter yet another benchmark. Where  = 5, 2 = 6/(5 – 4) = 6. At the other extreme,  = ∞ generates a Gaussian distribution, whose excess kurtosis is 0.

3.3. Performing Parametric VaR Analysis with Student’s t-Distribution

The following table reports parametric VaR values according to Student’s t-distribution for different values of . We return to our original assumption of a $1,000,000 portfolio with zero mean and unit variance (100 bps) on a Gaussian model. The quantile function for

Central Bank Journal of Law and Finance, No. 1/2015 13 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

varying levels of p, or zp, is the most salient difference, but not the only difference. Recall  that the standard deviation of Student’s t-distribution,   , will exceed the   2 corresponding Gaussian value except in the special, trivial case where  = ∞. Recall further

that VaR for a single time period is the product of the distribution-specific quantile (zp), the standard deviation of the distribution (), and the value of the portfolio: VaR  z    v  p p (taking care to distinguish the Roman letter v as the variable for the value of the portfolio from the Greek letter  as the variable designating the number of degrees of freedom in Student’s t-distribution). As a result, in stark contrast with the table describing Gaussian  VaR, values for VaRp in the following table may not be derived simply by multiplying

$10,000 by –zp.

Table 2

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

 = 4, 2 = ∞ –7.1732 –4.6041 –3.7470 –2.7765 –2.1319 –1.5332

  2 , zp =

VaRp $101,444 $65,112 $52,990 $39,265 $30,149 $21,683

  = 4.1, 2 = 60 –6.9986 –4.5283 –3.6970 –2.7499 –2.1170 –1.5260

4.1   2.1 , zp =

VaRp $97,790 $63,273 $51,567 $38,424 $29,580 $21,323

  = 5, 2 = 6 –5.8934 –4.0321 –3.3649 –2.5706 –2.0151 –1.4759

5   3 , zp =

VaRp $76,084 $52,055 $43,441 $33,186 $26,014 $19,054

  = 6, 2 = 3 –5.2076 –3.7074 –3.1427 –2.4469 –1.9432 –1.4398

14 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

3   2 , zp =

VaRp $63,780 $45,407 $38,490 $29,968 $23,799 $17,633

  = 7, 2 = 2 –4.7853 –3.4995 –2.9980 –2.3646 –1.8946 –1.4149

7   5 , zp =

VaRp $56,620 $41,406 $35,472 $27,979 $22,417 $16,742

  = 9, 2 = 1.2 –4.2969 –3.2498 –2.8214 –2.2622 –1.8331 –1.3830

9   7 , zp =

VaRp $48,722 $36,850 $31,992 $25,650 $20,786 $15,682

  = 10, 2 = 1 –4.1437 –3.1623 –2.7638 –2.2281 –1.8125 –1.3722

5   4 , zp =

VaRp $46,328 $35,434 $30,900 $24,911 $20,264 $15,341

  = 12, 2 = ¾ –3.9296 –3.0545 –2.6810 –2.1988 –1.7823 –1.3566

6   5 , zp =

VaRp $43,047 $33,461 $29,369 $24,087 $19,524 $14,861

  = 16, 2 = ½ –3.6862 –2.9208 –2.5835 –2.1191 –1.7459 –1.3368

8   7 , zp =

VaRp $39,407 $31,224 $27,619 $22,654 $18,664 $14,291



Central Bank Journal of Law and Finance, No. 1/2015 15 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

 = ∞, 2 = 0 –3.0902 –2.5758 –2.3264 –1.9597 –1.6449 –1.2816

 1, zp =

VaRp $30,902 $25,758 $23,264 $19,597 $16,449 $12,816  As Table 2 illustrates, calibrating parametric VaR according to , the number of degrees of freedom in Student’s t-distribution, generates a wide range of VaR values for different confidence levels represented by the variable p. For each value of , there is a corresponding 6 excess kurtosis that is very easy to calculate:   . As a quantile-based risk measure, 2   4 VaR is elicitable.90 This is the technique’s primary methodological virtue. Elicitability from historic data means that VaR can be subjected to a consistent scoring function that properly reports the measure’s reliability in forecasting future losses.91 In practical terms, this  property invites us to set the value of  according to observed levels of kurtosis in previous

returns. Subject to the constraints  > 4 and  2  0, rearrangement of the formula for excess kurtosis in Student’s t-distribution yields an extremely simple formula for  in terms of the 6 kurtosis observed in the distribution of previous returns:    4 .   2 Choosing the appropriate value for the  parameter, or the number of degrees of freedom, in VaR analysis using Student’s t-distribution therefore hinges, at least in the first instance, on historical levels of kurtosis exhibited by returns on the relevant asset class. Consider the following summary of historical descriptive statistics for thirteen noncash asset classes from February 1990 through May 2010:92

Table 3

Implied degrees of Excess kurtosis: Asset class Kurtosis (K) freedom:  K  3 6 2    4  2

Large value 5.06 2.06 6.91

Large growth 4.19 1.19 9.04  Small value 5.16 2.16 6.78

16 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

Implied degrees of Excess kurtosis: Asset class Kurtosis (K) freedom:  K 3 6 2    4  2

Small growth 3.84 0.84 11.14

Non-U.S. developed 4.29 1.29 8.65  markets

Emerging markets 4.72 1.72 7.49

Commodities 6.67 3.67 5.63

Non-U.S. REITs 5.03 2.03 6.96

U.S. REITs 10.51 7.51 4.80

U.S. TIPs 8.27 5.27 5.14

Other U.S. bonds 3.72 0.72 12.33

Non-U.S. bonds 3.54 0.54 15.11

Global high-yield 12.50 9.50 4.63 bonds Calculating VaR in a portfolio consisting of one of these asset classes is straightforward. The number of degrees of freedom in Student’s t-distribution is a simple function of historically 6 observed kurtosis:    4. Intuitively, lower levels of kurtosis correspond to a K  3 greater degrees of freedom. Since Student’s t-distribution converges with the Gaussian distribution as  , this model hews more closely to the Gaussian baseline for returns exhibiting low kurtosis. 



Central Bank Journal of Law and Finance, No. 1/2015 17 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

4. COMPARING STUDENT’S T-DISTRIBUTION WITH THE LOGISTIC DISTRIBUTION

Bell curves come in different configurations. Like the family of multivariate Student’s t- distributions, the entire family of multivariate logistic distributions belongs to the same class of jointly elliptical distributions that includes the standard normal distribution.93 The simplest version of the logistic distribution94 provides an instructive contrast with the t- distributions we have used thus far to enhance the robustness of our parametric VaR analysis, relative to the Gaussian baseline. Like Student’s t-distribution, the logistic distribution has heavier tails than the Gaussian distribution:

Figure 2

Source: http://ars.els-cdn.com/content/image/1-s2.0-S0360544210000617-gr24.jpg Unlike the Gaussian distribution or Student’s t-distribution, the logistic distribution has an inverse cumulative distribution function with an easily analyzed, closed-form solution:95

 p  F 1(p; , s)    sln  1 p

F−1(p) is the logistic distribution’s analog to the notation for the inverse cumulative distribution function of the normal distribution, –1(p). As always, μ represents the mean. s  is not quite standard deviation σ, but rather a parameter proportional to it. The ratio of s to σ can be calculated from the variance of the logistic function (which, as with other distributions, is the square of standard deviation σ):96

18 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

 2 2  s2 3    s 3 3 s  

The ratio √3/π is approximately 0.5513.

 The same result may be derived from the simple form of the moment-generating function for the logistic distribution:

n n n n E[( X  ) ]  s  (2  2)  Bn

97 where Bn indicates the nth Bernoulli number. Substituting the values s  3 , B  1 , B   1 yields these values for the variance, standard deviation, and   2 6 4 30 excess kurtosis of a logistic distribution:98

2   1

 6  2  5 1.2

Student’s t-distribution where  = 9 has the same excess kurtosis as the logistic distribution 6 whose s scale parameter is √3/π. For both distributions,  2  5 . But since these two  distributions have rather distinctive shapes, and since the logistic distribution maintains unit variance, while Student’s t-distribution with 9 degrees of freedom exhibits a variance of 9/7 and a standard deviation of 3/√7 (approximately 1.1339), the two distributions yield different  VaRp values despite sharing the same excess kurtosis:

Table 4

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

Logistic: –3.8079 –2.9184 –2.5334 –2.0198 –1.6234 –1.2114 3 s   , 2 = 1.2

 1, zp = 

 Central Bank Journal of Law and Finance, No. 1/2015 19 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

VaRp $38,079 $29,184 $25,334 $20,198 $16,234 $12,114

 = 9, 2 = 1.2 –4.2969 –3.2498 –2.8214 –2.2622 –1.8331 –1.3830

9   7 , zp =

VaRp $48,722 $36,850 $31,992 $25,650 $20,786 $15,682

 Indeed, the VaR values for p > 0.025 in the logistic model are lower than their corresponding values under the Gaussian model (reported above as Student’s t-distribution with infinite degrees of freedom). Casual observation suggests that the logistic model, despite its excess kurtosis of 1.2, is more closely aligned with models based on Student’s t-distributions with 12 or even 16 degrees of freedom. To see how and why Student’s t-distribution with 9 degrees of freedom does not align with a logistic distribution whose scale parameter s is √3/π, even though both distributions share

the same level of excess kurtosis (2 = 1.2), we need to specify Student’s t-distribution for  = 9. Consider the following restatement of the probability density function for Student’s t- distribution:

 1 ( 2 ) f (t)  2  1  t 2  ( 2) (1  )

2 9 3 6 Substituting  = 9 so that   7 ,   7 ,  2  5 produces the following probability density function:  (5) f (t)  2 9 1t 5 3  ( 2)( 9 )

Further simplification requires application of the gamma function and rearrangement of one of the terms in the denominator: 

20 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

(5)  4! 24 9 105 ( 2)  16  5 1 t 2  (9  t 2 )5     9  95

Substitution of these values yields a tractable version of the probability density function in simple analytical terms:

 10 7 9 4! 8 16 3 2  3 f (t)  2 5   3  105  (9t ) 105 (9  t 2)5 5 7  (9  t 2)5 16 95

Plots of the Gaussian distribution, the logistic distribution, the Student’s t-distribution at   = 9 for –4 < t < 0 show how these three curves diverge:99

Figure 3

Even more striking are plots for all three distributions at –6 < t < –2:100

Central Bank Journal of Law and Finance, No. 1/2015 21 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

Figure 4 This final plot shows how Student’s t-distribution with 9 degrees of freedom has a discernibly fatter tail than the logistic distribution with scale parameter s = 3/√π, even though both distributions exhibit the same kurtosis, and how both of those distributions exhibit kurtosis in excess of the Gaussian distribution. In parametric VaR, or any other application exploiting the greater robustness of Student’s t- distribution, the number of degrees of freedom, , not only controls the kurtosis of the distribution, but also influences the distribution’s overall scale and dispersion.101 At every confidence level, the quantile component of Student’s t-distribution with 9 degrees of freedom exceeds that of logistic VaR by roughly 3/√7, the standard deviation of Student’s t- distribution where  = 9. And since VaR requires the multiplication by the standard deviation, the ultimate difference between VaR according to Student’s t-distribution and logistic VaR is roughly 9/7, or the variance of Student’s t-distribution with 9 degrees of freedom. One further implication of parametric modeling of market returns is that leptokurtic “fat tail” behavior may arise from the heteroskedasticity of the returns. Volatility begets volatility, and sharp deviations from expected returns are disproportionately likely to be followed by comparably sharp deviations.102 Serially correlated returns should be expected to exhibit greater variance and greater kurtosis.103 Even if returns do converge over the long run in accordance with some variant of the central limit theorem, modeling returns according to Student’s t-distribution with a sufficiently high number of degrees of freedom guarantees that both variance and kurtosis will be determinate, despite the failure of serially correlated returns to be independent and identically distributed.

22 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

Excess kurtosis in the logistic distribution, as it turns out, has its greatest impact as the confidence level and the magnitude of a feared loss move further toward the negative tail of the distribution. Even the mildly leptokurtic logistic model demands increasingly stringent levels of caution as we move into that distribution’s relatively fat tails. At modest levels such as p = 0.01, the substitution of a logistic model for a normal model results in an 8.9 percent increase in VaR. At p = 0.001, however, the difference arising from the substitution of the logistic model becomes 23.2 percent. The implementation of Student’s t-distribution with 9 degrees of freedom with the same excess kurtosis of 1.2, raises VaR for p = 0.001 by 57.7 percent, relative to the Gaussian baseline for that confidence level. Either 23.2 or 57.7 percent represents a steep increase in the amount of risk that a portfolio manager must hedge, either through diversification of assets or (more likely under the demanding regulatory stress tests of Basel 2.5 and other similar sources of law) through aggressive reduction in the overall size of the portfolio. Greater caution against heightened forecasts of risk is precisely what many bankers want to avoid. Caution in whatever form — whether diversification, portfolio reduction, or hedging through futures and options — always comes at a price. Recall also that we have compressed the time horizon for all of our parametric VaR calculations to a single day. A longer time horizon expands the gap between Gaussian VaR and its relatively leptokurtic analogs, and along with that gap, the portfolio manager’s reluctance to adopt a measure of risk that can be reliably expected to reduce profitability. Over any time horizon, the risk of multiple occurrences within a single test period also compounds the propensity to underestimate aggregate losses.104 What we should expect to observe in real-world regulatory settings is a systematic failure to implement sufficient safeguards against market risk to the balance sheets of individual investors, of nonprofit endowments, and of systemically important financial institutions.

5. EXPECTED SHORTFALL AS A RESPONSE TO THE MODEL RISK OF PARAMETRIC VAR

5.1. Value-at-Risk Versus Expected Shortfall

Reducing the vulnerability of parametric VaR to model risk by improving its robustness addresses merely one threat to the reliability of value-at-risk analysis. The most serious menace to VaR — and to the technique’s viability as the Basel accords’ preferred approach to financial risk management — lies in VaR’s failure to does not satisfy the theoretical rigors demanded of “coherent” measures of risk.105 The expected shortfall for any confidence interval, which is derived directly from VaR for that interval, is subadditive and coherent,

Central Bank Journal of Law and Finance, No. 1/2015 23 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

while VaR is not.106 The allure of subadditivity and coherence supports Basel III’s embrace of expected shortfall as the international banking system’s preferred measure of market risk. The overriding allure of VaR — that it reports risk as a single number, either a percentage or a fixed dollar amount — invites reliance on a measure that may in fact severely underestimate tail risk. VaR relies on the greatest lower bound on an arbitrarily defined risk frontier over an arbitrarily fixed period of time. Because Basel II’s measurement of VaR at a 99 percent confidence interval over a holding period of ten trading days relies on a simple quantile analysis,107 this simpler approach to risk management necessarily disregards the magnitude and distribution of risks in the tail beyond the designated quantile boundary:108 A risk measure can be characterized by the weights its assigns to quantiles of the loss distribution. VaR gives a 100% weight to the Xth quantile and zero to other quantiles. Expected shortfall gives equal weight to all quantiles greater than the Xth quantile and zero weight to all quantiles below the Xth quantile. … [A] spectral risk measure is coherent (i.e., it satisfies the subadditivity condition) if the weight assigned to the qth quantile of the loss distribution is a nondecreasing function of q. Expected shortfall satisfies this condition. VaR does not, because the weights assigned to quantiles greater than X are less than the weight assigned to the Xth quantile.109

Gaussian VaR, it should be noted, is coherent. The subadditivity of normally distributed value-at-risk can be analytically proven:110

2 2 xy  x y  2(x,y) (x,y) (x,y)  xy 2 2 xy  x y  2(x,y)  xy 2 2 xy  x y  2xy, Q 1 (x,y) 1 2 xy  (x y )

xy  x y

Outside special cases such as this, however, VaR is not invariably subadditive and coherent.111

 5.2. Extrapolating Expected Shortfall from Parametric VaR One possible response to the incoherence of VaR is to substitute the corresponding value for expected shortfall for any parametrically computed amount of value-at-risk. For each point

24 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

estimate of VaR based on mathematically defined mean, variance, kurtosis, and confidence level, there is a corresponding level of expected shortfall.112 The substitution of expected shortfall for VaR therefore ameliorates model risk. In circumstances such as those presented by this article, which do not involve the computation of expected shortfall according to skewness as well as kurtosis and confidence level (a “complicated” process that “generally has no closed-form solution”),113 the cumulative distribution function for Student’s t-distribution provides point estimates for expected shortfall corresponding to VaR at critical confidence levels such as 0.001, 0.005, and 0.01.

In its most general form, expected shortfall, also known as conditional value-at-risk or expected tail loss, is the average of all losses equal to or worse than VaR at a particular confidence level.114 Formally, for a continuous distribution where L represents the loss function and p represents the critical confidence level, the following equation describes expected shortfall:

1 p ESp (L)  E(L | L VaRp )   VaRx (L) dx p 0

This abstract expression of expected shortfall may be more easily understood as a worked example. We return to the original specification of parametric VaR according to Student’s t-  distribution for some real, positive number of degrees of freedom calibrated according to the

variance and kurtosis of historic returns. For a single time interval, VaRp  zp    v. It follows that expected shortfall at the same confidence level p is the definite integral of this formula over the interval 0 to p, multiplied by the reciprocal of p:

1 p 1 p  ESp (L)   VaRx (L) dx   zp    v dx p 0 p 0

Let us once again set v at $1 million. We assume a confidence level p of 0.01. For purposes of illustration, we have presumably conducted parametric VaR analysis according to  Student’s t-distribution with 6 degrees of freedom, a curve whose variance is 3/2 and whose 3 excess kurtosis is 3. Formally,   6;   2 ;  2  3. The crucial value in the definite

integral that will report expected shortfall for this distribution at p = 0.01 is the integral of zp, which in turn is the integral of the inverse cumulative distribution function of Student’s t- distribution. Recall the formula for the cumulative distribution function of Student’s t-  distribution. For confidence intervals less than ½:

Central Bank Journal of Law and Finance, No. 1/2015 25 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

1 1 zp  F (x)    1  1 1 I2x ( 2 , 2)

1 For known values of  and p, the indefinite integral of – F (x) can be expressed in terms of the inverse regularized beta function. For  = 6:115 

1 1 1 3  1 1 1  I2x (3, 2 ) 1 1 3 I2x (3, 2)   F6 (x) dx   6 1 1 dx  6   C I ( 6 , 1 ) 16 1 1 2x 2 2 1 I2x (3, 2)

The value of this definite integral for p = 0 to 0.01 is approximately –0.0403253:116

0.01  1  6 1 6 1 1 dx  0.0403253 0 I2x ( 2 , 2)

Dividing this figure by p = 0.01 reports a value that is directly comparable to –zp in

conventional VaR analysis. Whereas –zp for p = 0.01 and  = 6 is approximately 3.1427,  p 1 z d for those parameters is approximately 4.03253. Multiplying this figure by the p 0 p 3 standard deviation ( 2 ), the value of baseline unit variance (i.e., 100 bps), and the value of the portfolio ($1 million) yields an expected shortfall of $49,388 for the parameters

 3   6;   2 ;  2  3. This risk statistic reports a considerably larger value than parametric VaR according to Student’s t-distribution for  = 6 and p = 0.01, which was $38,490.  The following table reports expected shortfall corresponding to parametric VaR values for representative numbers of degrees of freedom (4, 6, 10, and 16):

Table 5

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

 = 4, 2 = ∞ –7.1732 –4.6041 –3.7470 –2.7765 –2.1319 –1.5332

  2 , zp =



26 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

VaRp $101,444 $65,112 $52,990 $39,265 $30,149 $21,683

p –9.6862 –6.3248 –5.2206 –3.9936 –3.2029 –2.4993 1 z d p 0 p

ESp $136,984 $89,447 $73,830 $56,477 $45,296 $35,346   = 6, 2 = 3 –5.2076 –3.7074 –3.1427 –2.4469 –1.9432 –1.4398

3   2 , zp =

VaRp $63,780 $45,407 $38,490 $29,968 $23,799 $17,633

 –6.4158 –4.6756 –4.0325 –3.2562 –2.7107 –2.1872

ESp $78,577 $57,265 $49,388 $39,880 $33,200 $26,787

 = 10, 2 = 1 –4.1437 –3.1623 –2.7638 –2.2281 –1.8125 –1.3722

5   4 , zp =

VaRp $46,328 $35,434 $30,900 $24,911 $20,264 $15,341

 –4.8129 –3.7836 –3.3633 –2.8190 –2.4084 –1.9892

ESp $53,810 $42,301 $37,602 $31,517 $26,927 $22,240

 = 16, 2 = ½ –3.6862 –2.9208 –2.5835 –2.1191 –1.7459 –1.3368

8   7 , zp =

VaRp $39,407 $31,224 $27,619 $22,654 $18,664 $14,291



Central Bank Journal of Law and Finance, No. 1/2015 27 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

p 0.1% 0.5% 1.0% 2.5% 5.0% 10%

p –4.1660 –3.3980 –3.0658 –2.6177 –2.2657 –1.8937 1 z d p 0 p

ESp $44,536 $36,327 $32,774 $27,984 $24,221 $20,244  5.3. Expected Shortfall in a Parametric Model Based on the Logistic Distribution

If only in the interest of completeness, I shall specify the calculation of expected shortfall corresponding to different confidence levels for parametric VaR based on the logistic distribution. This exercise begins with an application to the logistic distribution of the general formula for expected shortfall as a function of the corresponding form of parametric VaR:

p p 1 1 1 ESp (L)   VaRx (L) dx   Fp (x)    v dx p 0 p 0

Because the logistic function can be specified through elementary functions, the indefinite and definite integrals of its inverse cumulative distribution function are straightforward:  3  x  F 1(x)   ln   1 x 

1 3 x  F (x) dx   [ln(1 x)  x ln( 1x )] C p p F 1(p) dx  3ln( p ) dx 0 0  1 p

p p 0.01 p For p = 0.01, 1  3ln( ) dx  1  3ln( ) dx  3.08753. This value lies p 0  1 p 0.01 0  1 p  further from zero than the inverse cumulative distribution function for the logistic distribution where p = 0.01. Namely, z0.01 ≈ –2.5334. The following table illustrates the difference between VaR and expected shortfall based on a logistic distribution for a range of confidence intervals:

Table 6

P 0.1% 0.5% 1.0% 2.5% 5.0% 10%

28 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

P 0.1% 0.5% 1.0% 2.5% 5.0% 10%

Logistic: –3.8079 –2.9184 –2.5334 –2.0198 –1.6234 –1.2114 3 s   , 2 = 1.2

 1, zp =

 VaRp $38,079 $29,184 $25,334 $20,198 $16,234 $12,114

 p –4.3595 –3.4711 –3.0875 –2.5782 –2.1889 –1.7923 1 z d p 0 p

ESp $43,595 $34,711 $30,875 $25,782 $21,889 $17,923

 6. UNADDRESSED RISKS: ESTIMATION, BEHAVIOR, AND SYSTEMIC EFFECTS

Model risk, as demonstrated by the gap between VaR and its corresponding expected shortfall values, is hardly the only threat to proper financial risk assessment. Even if we have properly modeled risk, whether by engaging in thorough nonparametric VaR, by specifying the proper parameters in a more accurate parametric model of value at risk, or by substituting more conservative (and coherent) values for expected shortfall in place of VaR, we cannot eliminate the problem of straightforward mistakes in estimation.117 As the economic crisis of 2008-09 painfully demonstrated, VaR fares poorly during periods of market stress.118 As befits an econometric model whose roots assume a symmetrical distribution of returns, VaR also proves fragile during periods of prosperity. Indeed, the very elegance of VaR gives rise to a false sense of security. VaR is at once easy to misunderstand and dangerous when misunderstood.119 Leptokurtosis poses problems in both tails — that is, during crisis and during prosperity. Because tail risk, scientifically speaking, defies measurement, the mere presence of a concrete, quantified figure such as VaR invites risk- taking wholly unwarranted by the real but unknown (and unknowable) state of economic affairs. Basel 2.5 sought to enhance the robustness of VaR by requiring that this exercise in risk management to account for historic evidence of stressed market conditions. The ease with which VaR is implemented and the illusory comfort provided by a formulaic measure of risk pose behavioral risks of their own. Psychological anchors provided by VaR backfire when risks within the unobserved tail eventually materialize. In 2009 testimony urging Congress to ban VaR, Nassim Nicholas Taleb criticized both the scientific uncertainty of VaR and its psychological effects on traders.120 Within the financial industry itself, David

Central Bank Journal of Law and Finance, No. 1/2015 29 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

Einhorn of Greenlight Capital has echoed these criticisms, alleging that VaR creates perverse incentives to take “excessive but remote risks” and is “potentially catastrophic when its use creates a false sense of security among senior executives and watchdogs.”121 Finally, the nearly global embrace of VaR analysis has resulted in an internally contradictory and potentially destructive “conflict — the uniform application of a risk measure that presumes independence and randomness.”122 This old adage bears remembering: In unruly markets, the only thing that rises is correlation.123 VaR’s troubles merely begin with its mistaken assumption that risk is randomly distributed, rather than correlated.124 The technique’s popularity among financial institutions and their regulators poses a deeper systemic threat. The very uniformity of a global financial system heeding the same rules may affirmatively raise correlations and systemic risk. By “promot[ing] coordination” within financial markets, uniformity in regulatory safeguards against market risk “can erode management tools premised on randomness and independent action and alter the dynamics that make risk management effective.”125 Legal coordination of markets, paradoxically and perversely, amplifies financial risks and deepens the scale of losses when they accrue.126 Indeed, the most highly rated assets, which effectively lower capital adequacy requirements, may expose banks to greater downside risk in stressed markets, since the systemic risk that inheres in highly rated assets is not the idiosyncratic, diversifiable risk associated with more speculative instruments.127 The trouble arises from rising correlation among asset classes during times of stress and from interconnectedness among banks and among their business partners. Flights to quality and resulting stress-driven changes in correlation among asset classes indicate conditions under which banks and other financial institutions will face heightened systemic risk.128 Measures of systemic risk arise from the insight that past data on volatility and correlation, having become procyclical, are less valuable than future forecasts of risk adjusted for economic cycles and anticipated mismatches between liquidity needs and the weighted maturity of individual banks’ balance sheets.129 VaR and expected shortfall, respectively, have inspired competing measures of systemic risk: CoVar (value-at-risk of the financial system, conditional on banks being in stress)130 and SRISK, a conditional capital shortfall index that takes account of each bank’s “systemic expected shortfall,” or “its propensity to be undercapitalized when the [overall financial] system as a whole is undercapitalized.”131 The cataclysmic prospect of global, systemic bank failure lies beyond the scope of this article. I nevertheless offer one grim observation on the prospects for regulatory success. Even the more modest enterprise of microprudential risk management, one bank at a time, may exceed the capacity of existing quantitative tools. In response to the methodological limitations on risk management, the Basel accords and other sources of international financial law often apply arbitrary multipliers such as three. For instance, one model for

30 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen improving the robustness of VaR prescribes three states of preparedness based on multiples of VaR: 1 through 3 for normal conditions, 3 through 10 for stress testing, and 10+ for all other means of hedging or insuring against contingencies beyond realistic business planning.132 For its part, this article has proposed readily implemented parametric alternatives to conventional Gaussian value-at-risk analysis. I have proposed calibrating , the number of degrees of freedom in Student’s t-distribution, according to the observed levels of kurtosis in the historic distribution of returns. The road of excess kurtosis leads to the palace of financial wisdom.133 But there may be no quantifiably reliable way of adopting a sufficiently conservative model of risk. No historic model of economic risk can predict extreme tail events. If we contemplate the sort of black swan event that arrives in one out of a million trading days, roughly once every 4,000 years, the difference becomes even more dramatic.134 A Gaussian model evaluates VaR1e–6 as a 4.8σ event and contemplates $47,534 in value-at-risk from an event expected to occur only once since the Middle Kingdom ruled Egypt. But the logistic model calculates VaR1e–6 as $76,169, while the application of Student’s t-distribution with 9 degrees of freedom (corresponding to the logistic model’s excess kurtosis of 1.2) would put VaR1e–6 at $121,556. Gaussian VaR would evaluate potential losses of $76,169 and $121,556, respectively, as events roughly equal to 7.6σ and 12σ. Ever deeper risks prompt a quest for ever higher confidence levels and more stringent capital reserve requirements. The eventual failure of such measures exposes our “lack [of] good analytic techniques for quantifying total risk when the distribution has a fat tail.”135 Although adopting a more leptokurtic statistical distribution improves the robustness of our exercise in parametric VaR analysis, we have strong reason to distrust all historic models. Economic history is pockmarked with extreme tail events exceeding traditional forecasts. It bears remembering that the record of monthly fluctuations in American stock market prices from 1871 through 2010 has reported 10σ events in both directions,136 even though we have already surmised (quite erroneously) that a 5σ event happens once every 4,000 years. As a first response, it may be more realistic — and perhaps even feasible — for rough forecasting techniques such as parametric VaR to anticipate events occurring once every 125 years or so. An easily remembered benchmark is once in 215 (32,768) trading days, or roughly 130 years containing 252 trading days. That time horizon predates not only the passage of modern securities laws in the United States by half a century;137 it also precedes the passage of comprehensive corporation laws in New Jersey and Delaware, the states that pioneered modern corporate law.138 The following table describes how each of the parametric VaR and expected shortfall models in this article would handle a confidence level of p = 2–15:

Central Bank Journal of Law and Finance, No. 1/2015 31 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

Table 7

p Degrees of z 15 VaR 15 ES 15 2  2 2 1 z d 2 freedom () p 0 p

4 ∞ 2 –17.6126 $249,080 –23.5338 $332,818 5 6 5  –12.3755 $159,767 –15.5411 $200,634 3  6 3 3 –9.90851 $121,354 –11.9801 $146,725  2 7 2 7 –8.51579 $100,760 –10.0400 $118,795  5 9 1.2 9 –7.03231 $79,739 –8.03907 $91,155  7 Logistic 1.2 1 –5.73228 $57,323 –6.28359 $62,836  10 1 5 –6.59669 $73,753 –7.46606 $83,473 4  12 ¾ 6 –6.01199 $65,858 –6.70853 $73,488 5

16 ½ 8 –5.38245 $57,541 –5.90918 $63,172  7 ∞: Gaussian 0 1 –4.00877 $40,088 –4.23399 $42,340 

Notably, the most conservative value for VaR 15 in the foregoing table provides an  2 intuitively plausible estimate of the single largest daily drop in the 119-year history of the Dow Jones Industrial Average: the 22.61 percent drop on Black Monday, October 19, 1987. Just as notably, the VaR forecast that comes closest to matching that loss is the forecast with  infinite kurtosis. A 22.61 percent drop in value falls roughly between expected shortfall forecasts for 4 and 4.5 degrees of freedom. At  = 5, expected shortfall is slightly over $200,000. At  = 4.5, expected shortfall falls just under $250,000. The corresponding values for excess kurtosis would fall between 6 (at  = 5) and 12 (at  = 4.5). These values fall on either side of the highest level of excess kurtosis observed in one study of returns on publicly traded asset classes from 1990 through 2010.139 Lest we vest too much trust in the modeling and forecasting prowess of banks and their regulators, financial history suggests that peacetime market collapses may represent only a small portion of risk that remains unobservable and therefore beyond the theoretical reach of even the most thorough models. September 1939, the opening month of World War II, recorded a 9σ departure from the history of correlations between market returns and beta from 1955 through 1968.140 Smaller, less liquid markets reveal even greater shocks. A 50-

32 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

year survey of oil prices, from 1960 through 2010, has revealed a 37σ event in 1973.141 37σ! It would not have been unreasonable to surmise that “the economic world as we knew it was coming to an end.”142 The expected time between 25 events — roughly 1.309 x 10136 years — spans “more millennia than the universe has particles.”143 Relying on “VaR models … to explain … 25-standard deviation moves, day after day,” is not just wrong. It is “catastrophically wrong.”144 Loss scenarios of this magnitude force us to contemplate appropriate responses to “infinite disutility.”145 The presence of fat tails, especially when they are fat enough to push probability distributions toward the intractable extreme of infinite variance, thrusts us into “ignorance about both the exact form of the distribution (e.g., normal, Pareto, or exponential) and the exact parameters of the distribution.”146 The future of financial risk management is a future of staggering data with diminished confidence in beautiful but deceptive mathematical models. We may anticipate a parade of imperfectly articulated hypotheses and efforts at verification or falsification that will likely lead neither to elegant closed-form solutions nor to pathological functions. Blinded (as it were) by the epistemic limitations stemming from leptokurtosis, we invariably guess. And guesses, by their very nature, are often wrong. “Every year, if not every day, we have to wager our salvation upon some prophecy based upon imperfect knowledge.”147 Forecasting potential market losses according to levels of kurtosis approaching infinity reveals in concrete mathematical terms the extent of our ignorance.

Central Bank Journal of Law and Finance, No. 1/2015 33 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

NOTES

1 T.S. ELIOT, Little Gidding, in FOUR QUARTETS 49, 58 (Harcourt, Brace & Co. 1971; 1st ed. 1943). 2 See https://en.wikipedia.org/wiki/Method_of_moments_(statistics). 3 As this paper will demonstrate, the distribution and quantile functions of Student’s t-distribution may be expressed through the gamma and beta functions, which are interrelated special functions, and their generalizations, the incomplete gamma function and the incomplete beta function. 4 https://en.wikipedia.org/wiki/Kurtosis. 5 https://en.wikipedia.org/wiki/Kurtosis. 6 See Eugene F. Schuster, Classification of Probability Laws by Tail Behavior, 79 J. AM. STAT. ASS’N 936 (1984). 7 See generally Nicolas Champagnat, Madalina Deaconu, Antoine Lejay, Nicolas Navet & Souhail Boukherouaa, An Empirical Analysis of Heavy-Tails Behavior of Financial Data: The Case for Power Laws, https://hal.inria.fr/hal-00851429 (Aug. 14, 2013). 8 Martin L. Weitzman, A Review of The Stern Review on the Economics of Climate Change, 45 J. ECON. LIT. 703, 723 (2007). 9 Daniel A. Farber, Uncertainty, 99 GEO. L.J. 901, 926 (2011). 10 Id. 11 Id. at 925. 12 This is a distinction originating in FRANK KNIGHT, RISK, UNCERTAINTY, AND PROFIT 231-35 (1921); accord, e.g., Farber, Uncertainty, supra note 9, at 903 & n.5. 13 Farber, Uncertainty, supra note 9, at 906. 14 See generally LINDA ALLEN, JACOB BOUDOUKH & ANTHONY SAUNDERS, UNDERSTANDING MARKET, CREDIT, AND OPERATIONAL RISK: THE VALUE AT RISK APPROACH 1-20 (2004); SIMON BENNINGA & ZVI WIENER, VALUE AT RISK (1998); PHILIPPE JORION, VALUE AT RISK: THE NEW BENCHMARK FOR MANAGING FINANCIAL RISK (3d ed. 2006); Winfried G. Hallerbach, Decomposing Portfolio Value-at-Risk: A General Analysis, 5:2 RISK 1 (Winter 2002). 15 See generally James Ming Chen, Measuring Market Risk Under the Basel Accords: VaR, Stressed VaR, and Expected Shortfall, 8 AESTIMATIO 184 (2014) (comparing VaR, as endorsed by the Basel II accord, with stressed VaR and expected shortfalls, the risk measures endorsed, respectively, by Basel 2.5 and Basel III). The Basel accords are produced by the Basel Committee on Banking Supervision and have been published under these titles: (1) BASEL II: INTERNATIONAL CONVERGENCE OF CAPITAL MEASUREMENT AND CAPITAL STANDARDS: A REVISED FRAMEWORK (2006) (available at http://www.bis.org/publ/bcbs128.pdf); (2) REVISIONS TO THE BASEL II MARKET RISK FRAMEWORK (2009) (“Basel 2.5”) (available at http://www.bis.org/publ/bcbs158.pdf); (3) A GLOBAL RGULATORY FRAMEWORK FOR MORE RESILIENT BANKS AND BANKING SYSTEMS (2011) (“Basel III”) (available at http://www.bis.org/publ/bcbs189.pdf). 16 See BASEL II: INTERNATIONAL CONVERGENCE OF CAPITAL MEASUREMENT AND CAPITAL STANDARDS: A REVISED FRAMEWORK, part 2, ¶¶ 178-181 (June 2004) (available at http://www.bis.org/publ/bcbs107.pdf and http://www.bis.org/publ/bcbs107b.pdf). See generally CHRISTOPHER J. BRUMMER, SOFT LAW AND THE GLOBAL FINANCIAL SYSTEM: RULE-MAKING IN THE TWENTY-FIRST CENTURY (2011), Jim Chen, Book Review, Soft Law and the Global Financial System: Rule-Making in the Twenty-First Century, 25 EMORY INT’L L. REV. 1561 (2011).

34 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

17 See Charles K. Whitehead, Destructive Coordination, 96 CORNELL L. REV. 323, 343-44 (2011). 18 See, e.g., Risk-Based Capital Standards: Market Risk, 64 Fed. Reg. 19,034, 19,035 (April 19, 1999) (codified at 12 C.F.R. pt. 325); SEC Accounting Policies for Certain Derivative Instruments, 17 C.F.R. § 210.4-08(n); SEC Financial Statements, 17 C.F.R. § 228.310 (requiring firms to disclose information on derivative holdings through VaR-style models); European Commission, Internal Markets & Services DG, “Solvency II”: Frequently Asked Questions (available at http://ec.europa.eu/insternal_market/insurance/docs/solvency/solvency2/faq_en.pdf); BASEL COMMITTEE ON BANKING SUPERVISION, PROGRESS REPORT ON BASEL III IMPLEMENTATION (Oct. 2012) (available at http://www.bis.org/publ/bcbs232.pdf). See generally Whitehead, supra note 17, at 343- 44 (listing authorities in different jurisdictions that have adopted VaR for regulatory purposes). 19 See Whitehead, supra note 17, at 342. 20 Jonathan R. Macey, The Regulator Effect in Financial Regulation, 98 CORNELL L. REV. 591, 592-93 (2013). See generally Jeremy Berkowitz & James O’Brien, How Accurate Are Value-at-Risk Models at Commercial Banks?, 57 J. FIN. 1093 (2002). 21 See Douglas O. Edwards, Comment, An Unfortunate “Tail”: Reconsidering Risk Management Incentives After the Financial Crisis of 2007-2009, 81 COLO. L. REV. 247, 307 (2010). See generally Jorge Mina & Jerry Yi Xiao, Retern to RiskMetrics: The Evolution of a Standard (2001) (available at http://www.wu.ac.at/executiveeducation/institutes/banking/sbwl/lvs_ws/vk4/rrmfinal.pdf). 22 See KEVIN SHEPPARD, FINANCIAL ECONOMETRICS NOTES 494 (2012) (available at http://www.kevinsheppard.com/images/c/c0/Financial_Econometrics_2012-2013.pdf). 23 See, e.g., JOHN Y. CAMPBELL, ANDREW W. LO & A. CRAIG MACKINLAY, THE ECONOMETRICS OF FINANCIAL MARKETS 17, 81, 172, 498 (1997); Felipe M. Aparicio & Javier Estrada, Empirical Distributions of Stock Returns: European Securities Markets, 1990-95, 7 EUR. J. FIN. 1 (2001); Geert Bekaert, Claude Erb, Campbell R. Harvey & Tadas Viskanta, Distributional Characteristics of Emerging Market Returns and Asset Allocation, 24:2 J. PORTFOLIO MGMT. 102 (Winter 1998); Pornchai Chunhachinda, Krishnan Dandepani, Shahid Hamid & Arun J. Prakash, Portfolio Selection and Skewness: Evidence from International Stock Markets, 21 J. BANKING & FIN. 143 (1997). 24 See ALLEN, BOUDOUKH & SAUNDERS, supra note 14, at 8; JORION, supra note 14, at 110; ÖSTERREICHISCHE NATIONALBANK, FIVE GUIDELINES ON MARKET RISK: STRESS TESTING 3-4 (Wolfdietrich Grau ed., 1999). 25 See KEVIN DOWD, BEYOND VALUE AT RISK: THE NEW SCIENCE OF RISK MANAGEMENT 22 (1998). 26 LINDA ALLEN, JACOB BOUDOUKH & ANTHONY SAUNDERS, UNDERSTANDING MARKET, CREDIT, AND OPERATIONAL RISK: THE VALUE AT RISK APPROACH 8 (2004). 27 Id. 28 Elegant but arguably unrealistic financial theories built on “beautifully Platonic models on a Gaussian base,” NASSIM NICHOLAS TALEB, THE BLACK SWAN: THE IMPACT OF THE HIGHLY IMPROBABLE 279 (2007), include modern portfolio theory, see HARRY M. MARKOWITZ, PORTFOLIO SELECTION (1959); the capital asset pricing model, see Fischer Black, Capital Market Equilibrium with Restricted Borrowing, 45 J. BUS. 444 (1972); William F. Sharpe, Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of Risk, 19 J. FIN. 425 (1964); the Black-Scholes model of option pricing, see Fischer Black & Myron Scholes, The Pricing of Opetions and Corporate Liabilities, 81 J. POL. ECON. 637 (1973); Robert C. Merton, The Theory of Rational Option Pricing, 4 BELL J. ECON. 141 (1973); Merton’s distance-to- default model of credit risk, Robert C. Merton, On the Pricing of Corporate Debt: The Risk Structure of Interest Rates, 29 J. FIN. 449 (1974); and the Gaussian copula, see David X. Liu, On Default Correlation: A Copula Function Approach, 9:4 J. FIXED INCOME 43 (2000).

Central Bank Journal of Law and Finance, No. 1/2015 35 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

29 See, e.g., Szilard Páfka & Imre Kondor, Evaluating the RiskMetrics Methodology in Measuring Volatility and Value-at-Risk in Financial Markets, 299 PHYSICA A 305, 309 (2001). 30 See, e.g., PETER J. HUBER, ROBUST STATISTICS 1 (1981); Stephen Portnoy & Xuming He, A Robust Journey in the New Millennium, 95 J. AM. STAT. ASS’N 1331 (2000). 31 This example is drawn from ALLEN, BOUDOUKH & SAUNDERS, supra note 14, at 6-7. The ensuing discussion in text also draws upon (and considerably simplifies) the analysis outlined in JORION, supra note 14, at 106-13. 32 See Jón Daníelsson & Jean-Pierre Zigrand, On Time Scaling of Risk and the Square-Root-of-Time Rule, 30 J. BANKING & FIN. 2701, 2702 n.1 (2006). 33 See Johanna F. Ziegel, Coherence and Elicitability, at 1 (March 8, 2013) (available at http://arxiv.org/pdf/1303.1690v2) (forthcoming in MATHEMATICAL FINANCE). 34 See https://en.wikipedia.org/wiki/Quantile_function. 35 See https://en.wikipedia.org/wiki/Probit. 36 https://en.wikipedia.org/wiki/Normal_distribution. 37 Id. 38 See https://en.wikipedia.org/wiki/Standard_score. 39 See ALLEN, BOUDOUKH & SAUNDERS, supra note 14, at 7. 40 See http://en.wikipedia.org/wiki/Independent_and_identically_distributed_random_variables. 41 JORION, supra note 14, at 108. 42 See id. at 111 (equation 5.10). For a generalization of the square root rule for the addition of uncorrelated, independent and identically distributed variables to portfolio variance, hypothesis testing, and time series, see MICHAEL B. MILLER, MATHEMATICS AND STATISTICS FOR FINANCIAL RISK MANAGEMENT 45-46, 135-36, 215-16 (2d ed. 2014). 43 Compare Daniélsson & Zigrand, supra note 32 (demonstrating that scaling by the square root of time systematically underestimates risk, with greater downward bias as the time horizon increases, when the underlying risk factor follows a jump diffusion process) with Francis X. Diebold, Andrew Hickman, Atsushi Inoue & Til Schuermann, Scale Models, 11 RISK 104 (1998); Francis X. Diebold, Andrew Hickman, Atsushi Inoue & Til Schuermann, Converting 1-Day Volatility to h-Day Volatility: Scaling by √h Is Worse than You Think (July 3, 1997) (available at http://economics.sas.upenn.edu/~fdiebold/papers/paper18/dsi.pdf); Vikentia Provizionatou, Sheri Markose & Olaf Menkens, Empirical Scaling Rules for Value-at-Risk (VaR) (April 15, 2005) (available at http://web.econ.ku.dk/fru/conference/Programme/friday/a4/provizionatou_empirical%20scaling %20rule.pdf) (demonstrating that scaling by the square root of time systematically overestimates volatility over long time horizons when risk factors follow a GARCH(1,1) [generalized autoregressive conditional heteroskedasticity] process). The absence of “immediate alternatives to square-root of time scaling” has led the Basel Committee on Banking Supervision to acknowledge “the practical usefulness of square-root of time scaling” in spite of these theoretical limitations. Basel Committee on Banking Supervision, Messages from the Academic Literature on Risk Measurement for the Trading Book, at 8 (Jan. 31, 2011) (working paper no. 19) (available at http://www.bis.org/publ/bcbs_wp19.pdf). 44 See ALLEN, BOUDOUKH & SAUNDERS, supra note 14, at 7. 45 JORION, supra note 14, at 113. 46 Id.

36 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

47 See id. 48 See Gary Chamberlain, A Characterization of the Distributions That Imply Mean-Variance Utility Functions, 29 J. ECON. THEORY 185 (1983); Joel Owen & Ramon Rabinovitch, On the Class of Elliptical Distributions and Their Applications to the Theory of Portfolio Choice, 37 J. FIN. 745 (1983). 49 On the use of elliptical distributions to generalize the multivariate normal distribution, see generally KAI-TAI FANG, SAMUEL KOTZ & KAI WANG NG, SYMMETRIC MULTIVARIATE AND RELATED DISTRIBUTIONS (1990); ALEXANDER MCNEIL, RÜDIGER FREY & PAUL EMBRECHTS, QUANTITATIVE RISK MANAGEMENT: CONCEPTS, TECHNIQUES, AND TOOLS 72-78 (2005). 50 See generally PAUL LÉVY, CALCUL DES PROBABILITÉS (1925); Benoit Mandelbrot, The Pareto-Lévy Law and the Distribution of Income, 1 INT’L ECON. REV. 79 (1960); Benoit Mandelbrot, Stable Paretian Random Functions and the Multiplicative Variation of Income, 29 ECONOMETRICA 517 (1961). 51 James A. Xiong & Thomas M. Idzorek, The Impact of Skewness and Fat Tails on theAsset Allocation Decision, 67:2 FIN. ANALYSTS J. 23, 24 (March/April 2011). 52 See Benoit Mandelbrot, The Variation of Certain Speculative Prices, 36 J. BUS. 394 (1963). 53 See Eugene F. Fama, Mandelbrot and the Stable Paretian Hypothesis, 36 J. BUS. 420 (1963). 54 Eugene F. Fama, The Behavior of Stock-Market Prices, 38 J. BUS. 34, 42 (1965). 55 See Farber, Uncertainty, supra note 9, at 923-24 n.95 (explaining why the law of large numbers does not “shrink down” the sample mean of a Cauchy distribution even as sample size increases). 56 See Fama, The Behavior of Stock-Market Prices, supra note 54, at 43-45. 57 BENOIT B. MANDELBROT, THE FRACTAL GEOMETRY OF NATURE 337-38 (1983). 58 CAMPBELL, LO & MACKINLAY, supra note 23, at 18. 59 See https://en.wikipedia.org/wiki/Stable_distribution. 60 Xiong & Idzorek, supra note 51, at 24. 61 See JOHN P. NOLAN, STABLE DISTRIBUTIONS: MODELS FOR HEAVY TAILED DATA 13 (2009). 62 See https://en.wikipedia.org/wiki/Stable_distribution. 63 See Fama, The Behavior of Stock-Market Prices, supra note 54, at 102; https://en.wikipedia.org/wiki/Stable_distribution. 64 https://en.wikipedia.org/wiki/Cauchy_distribution. 65 CAMPBELL, LO & MACKINLAY, supra note 23, at 19. 66 Id. 67 See OLAV KALLENBERG, FOUNDATIONS OF MODERN PROBABILITY 66-67 (1997); C.C. Heyde, Central Limit Theorem, in 4 ENCYCLOPEDIA OF STATISTICAL SCIENCES 651 (Samuel Kotz et al. eds., 1983). On nonparametric methods for testing departures from the central limit theorem’s assumption of independent and identically distributed random variables, see generally SIDNEY SIEGEL & N. JOHN CASTELLAN, JR., NONPARAMETRIC STATISTICS FOR THE BEHAVIORAL SCIENCES (2d ed. 1988); A.N. Kolmogorov, Sulla determinazione empirica di unna legge di distribuzione, 4 G. IST. ITAL. ATTUARI 83 (1933); N. Smirnov, Table for Estimating the Goodness of Fit of Empirical Distributions, 19 ANNALS MATH. STAT. 279 (1948); William H. Kruskal & W. Allen Wallis, Use of Ranks in One-Criterion Variance Analysis, 47 J. AM. STAT. ASS’N 583 (1952); Robert G. Mogull, The One-Sample Runs Test: A Category of Exception, 19 J. EDUC. & BEHAV. STUD. 296 (1994). 68 CAMPBELL, LO & MACKINLAY, supra note 23, at 19.

Central Bank Journal of Law and Finance, No. 1/2015 37 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

69 JOHN MAYNARD KEYNES, A TRACT ON MONETARY REFORM 80 (1924); accord JOHN IRVING, THE WORLD ACCORDING TO GARP 688 (1998) (“[I]n the world according to Garp, we are all terminal cases.”) (1st ed. 1978). 70 CAMPBELL, LO & MACKINLAY, supra note 23, at 19. 71 See https://en.wikipedia.org/wiki/Student%27s_t-distribution. Unless indicated otherwise, subsequent discussions of Student’s t-distribution draw upon this source. The original paper describing the t-distribution was Student (William Sealy Gosset), The Probable Error of a Mean, 6 BIOMETRIKA 1 (1908). R.A. Fisher popularized Gosset’s work. See R.A. Fisher, Applications of Student’s” Distribution, 5 METRON 90 (1925). See generally Harold Hotelling, British Statistics and Statisticians Today, 25 J. AM. STAT. ASS’N 186 (1930). 72 Xiong & Idzorek, supra note 51, at 23. 73 See Zinoviy M. Landsman & Emiliano A. Valdez, Tail Conditional Expectations for Elliptical Distributions, 7:4 N. AM. ACTUARIAL J. 55, 59 (October 2003). 74 Such techniques are hardly limited to the use of Student’s t-distribution. See generally ERIC JONDEAU, SER-HUANG POON & MICHAEL ROCKINGER, FINANCIAL MODELING UNDER NON-GAUSSIAN DISTRIBUTIONS (2007). On the use of multiple Gaussian distributions to impart fatter tails and greater robustness to the modeling of stock market returns, see Peter K. Clark, A Subordinated Stochastic Process Model with Finite Variance for Speculative Prices, 41 ECONOMETRICA 135 (1973); N.E. Day, Estimating the Components of a Mixture of Normal Distributions, 56 BIOMETRIKA 463 (1969). 75 See Carmen Fernández & Mark F.J. Steel, On Bayesian Modeling of Fat Tails and Skewness, 93 J. AM. STAT. ASS’N 359 (1998). This technique has been applied to Student’s t-distribution. P. Lambert & S. Laurent, Modelling Financial Time Series Using GARCH-Type Models and a Skewed Student Density (2001) (working paper); accord RUEY S. TSAY, ANALYSIS OF FINANCIAL TIME SERIES § 3.4.3, at 122 (3d ed. 2010); Timotheos Angelidis, Alexandros Benos & Stavros Degiannakis, A Robust VaR Model Under Different Time Periods and Weighting Schemes, 28 REV. QUANT. FIN. & ACCOUNTING 187 (2007). 76 See J. AITCHISON & J.A.C. BROWN, THE LOGNORMAL DISTRIBUTION, WITH SPECIAL REFERENCE TO ITS USE IN ECONOMETRICS (1957); Hal Forsey, The Mathematician’s View: Modelling Uncertainty with the Three Parameter Lognormal, in MANAGING DOWNSIDE RISK IN FINANCIAL MARKETS 51 (Frank A. Sortino & Stephen E. Satchell eds., 2001); Eckhard Limpert, Werner A. Stahel & Markus Abbt, Log-Normal Distributions Across the Sciences: Keys and Clues, 51 BIOSCIENCE 341 (2001). 77 See Peter R. Fisk, The Graduation of Income Distributions, 29 ECONOMETRICA 171 (1961). 78 See N.L. Johnson, Systems of Frequency Curves Generated by Methods of Translation, 36 BIOMETRIKA 149 (1949). 79 See I.D. Hill, R. Hill & R.L. Holder, Fitting Johnson Curves by Moments, 25 APPLIED STAT. 180 (1976). 80 See Steven E. Posner & Moshe Arye Milevsky, Valuing Exotic Options by Approximating the SPD with Higher Moments, 7 J. FIN. ENG’G 109 (1998). 81 Robert C. Blattberg & Nicholas J. Gonedes, A Comparison of the Stable and Student Distributions as Statistical Models for Stock Prices, 47 J. BUS. 244, 252 (1974). 82 Id. 83 See Kenneth L. Lange, Roderick J.A. Little & Jeremy M.G. Taylor, Robust Statistical Modeling Using the t-Distribution, 84 J. AM. STAT. ASS’N 881 (1989). 84 See https://en.wikipedia.org/wiki/Gamma_function.

38 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

85 See https://en.wikipedia.org/wiki/Beta_function. 86 See https://en.wikipedia.org/wiki/Incomplete_beta_function. The beta and gamma functions are, respectively, Euler integrals of the first and second kind:

1 (x,y)  t x1(1 t)y1 dt 0  (z)  t z1e1 dt 0 (x)(y) (x,y)  (x  y)

See ALAN JEFFREY & HUI-HUI DAI, HANDBOOK OF MATHEMATICAL FORMULAS 234-35 (4th ed. 2008). As is evident from the foregoing identities, the beta function may be expressed in terms of the gamma function. The incomplete beta function is a generalization of the beta function, where the Euler integral of the first kind is calculated over the interval from 0 to a, such that the ordinary, complete beta function represents the special case where a = 1. 87 See http://www.wolframalpha.com/input/?i=inverse+cdf+student%27s+t+distribution; http://reference.wolfram.com/language/ref/InverseBetaRegularized.html. Wolfram’s documentation of the inverse cumulative distribution function of Student’s t-distribution stands in pleasant contrast with that of other software providers. The documentation for MATLAB, for instance, provides little beyond a verbal description of F–1(t) as “the solution of the cdf integral with parameter ν, where you supply the desired probability p.” http://www.mathworks.com/help/stats/tinv.html (documenting the function, x = tinv(p, nu)). 88 Blattberg & Goenedes, supra note 81, at 275. 89 Id.; see also supra text accompanying notes 65-70. 90 See Tilmann Gneiting, Making and Evaluating Point Forecasts, 106 J. AM. STAT. ASS’N 747, 766-68 (2011). For discussions of elicitability in the context of VaR and expected shortfall, see Chen, supra note 15, at 197-98; Ziegel, supra note 33. 91 See Gneiting, supra note 90, at 758-61. 92 See Xiong & Idzorek, supra note 51, at 29 (table 6) and 31 (figure 2). 93 See Landsman & Valdez, supra note 73, at 59-60. 94 See https://en.wikipedia.org/wiki/Logistic_distribution. Unless provided otherwise, further discussion of the logistic distribution is drawn from this source. 95 See https://en.wikipedia.org/wiki/Logistic_distribution#Quantile_function. 96 See https://en.wikipedia.org/wiki/Logistic_distribution. 97 See https://en.wikipedia.org/wiki/Bernoulli_number. 98 See https://en.wikipedia.org/wiki/Logistic_distribution. 99 See http://www.wolframalpha.com/input/?i=plot+pi*exp%28- pi*x%2Fsqrt%283%29%29%2Fsqrt%283%29%2F%28exp%28- pi*x%2Fsqrt%283%29%29%2B1%29%5E2+and+2519424%2F35%2Fpi%2F%28x%5E2%2B9%29% 5E5+and+exp%28-x%5E2%2F2%29%2Fsqrt%282*pi%29+for+x%3D-4+to+0&f=1. 100 See http://www.wolframalpha.com/input/?i=plot+pi*exp%28- pi*x%2Fsqrt%283%29%29%2Fsqrt%283%29%2F%28exp%28- pi*x%2Fsqrt%283%29%29%2B1%29%5E2+and+2519424%2F35%2Fpi%2F%28x%5E2%2B9%29% 5E5+and+exp%28-x%5E2%2F2%29%2Fsqrt%282*pi%29+for+x%3D-6+to+-2.

Central Bank Journal of Law and Finance, No. 1/2015 39 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

101 See ANDREW GELMAN, JOHN B. CARLIN, HAL S. STERN, DAVID B. DUNSON, AKI VEHTARI & DONALD B. RUBIN, BAYESIAN DATA ANALYSIS 68 (1995). 102 See, e.g., TSAY, supra note 75, §§ 3.4, 3.5, at 116, 132 (3d ed. 2010) (describing how conditional heteroskedastic time series models address volatility clustering in financial returns). 103 See Blattberg & Gonedes, supra note 81, at 275-76; Eugene F. Fama, The Behavior of Stock Market Prices, 38 J. BUS. 34 (1965). 104 See Daigee Shaw et al., A Probabilistic Seismic Risk Assessment of Building Losses in Taipei: An Application of HAZ-Taiwan with Its Pre-Processor and Post-Processor, 30 J. CHINESE INST. ENG’RS 279, 290 (2007). 105 See generally SHEPPARD, supra note 22, at 522-24; Phillipe Artzner, Freddy Delbaen, Jean-Marc Eber & David Heath, Coherent Measures of Risk, 9 MATH. FIN. 203 (1999). 106 See Chen, supra note 15; Ziegel, supra note 33. 107 See BASEL II, supra note 15, at 44. 108 See JOHN HULL, RISK MANAGEMENT AND FINANCIAL INSTITUTIONS 189 (3d ed. 2012). 109 Id. at 190. 110 See Ruey S. Tsay, Lecture Note of Bus 41202, Spring 2013: Value at Risk, Expected Shortfall & Risk Management, at 5 (2013) (available at http://faculty.chicagobooth.edu/ruey.tsay/teaching/bs41202/sp2013/lec9-13.pdf). 111 See, e.g. STUART A. KLUGMAN, HARRY H. PANJER & GORDON E. WILLMOT, LOSS MODELS: FROM DATA TO DECISIONS 44-45 (4th ed. 2012); Julia Lynn Wirch, Raising Value at Risk, 3 N. AM. ACTUARIAL J. 106 (1999). 112 See R. Tyrell Rockafellar & Stanislav Uryasev, Optimiziation of Conditional Value-at-Risk, 2:3 J. RISK 21 (Spring 2000). 113 Xiong & Idzorek, supra note 51, at 23. 114 See generally Carlo Acerbi & Dirk Tasche, Expected Shortfall: A Natural Coherent Alternative to Value at Risk, 31 ECON. NOTES 379 (2002); Basel Committee on Banking Supervision, Messages from the Academic Literature on Risk Management for the Trading Book (2011) (available at http://www.bis.org/publ/bcbs_wp19.pdf); Ziegel, supra note 33; https://en.wikipedia.org/wiki/Expected_shortfall. The information presented in this section is drawn from these sources. 115 To compute this indefinite integral, I enlisted Wolfram Alpha. See http://www.wolframalpha.com/input/?i=integrate+sqrt%286%29*sqrt%281%2Finversebetaregula rized%282*x%2C3%2C1%2F2%29-1%29; http://www.wolframalpha.com/input/?i=integrate+sqrt%281%2Finversebetaregularized%282*x% 2C3%2C1%2F2%29-1%29. 116 See http://www.wolframalpha.com/input/?i=integrate+sqrt%286%29*sqrt%281%2Finversebetaregula rized%282*x%2C3%2C1%2F2%29-1%29+for+x%3D0+to+.01. 117 See SHEPPARD, supra note 22, at 496. 118 See, e.g., Michael C. Macchiarola, Beware of Risk Everywhere: An Important Lesson from the Current Credit Crisis, 5 HASTINGS BUS. L.J. 267, 294-97 (2009). 119 See Joe Nocera, Risk Mismanagement, N.Y. TIMES MAGAZINE, Jan. 4, 2009.

40 Central Bank Journal of Law and Finance, No. 1/2015 James Ming Chen

120 See http://gop.science.house.gov/Media/hearings/oversight09/sept10/taleb.pdf. 121 David Einhorn, Private Profits and Socialized Risk, GARP RISK REV., June/July 2008, at 10. 122 Whitehead, supra note 17, at 346. 123 See, e.g., MARTIN L. LEIBOWITZ, ANTHONY BOVA & P. BRETT HAMMOND, THE ENDOWMENT MODEL OF INVESTING: RETURN, RISK, AND DIVERSIFICATION 235, 265 (2010) (observing that rising correlations in falling markets may imperil portfolios designed to weather variation and confer diversification under normal conditions); Malcolm P. Baker & Jeffrey Wurgler, Comovement and Predictable Relations Between Bonds and the Cross-Section of Stocks, 2 REV. ASSET PRICING STUD. 57 (2012); Rob Bauer, Roul Haerden & Roderick Molenaar, Asset Allocation in Stable and Unstable Times, 13:3 J. INVESTING 72 (Fall 2004); John Drzik, Richard J. Herring & Francis X. Diebold, The New Role of Risk Management: Rebuilding the Model, Knowledge@Wharton, http://knowledge.wharton.upenn.edu/article.cfm?articleid=2268 (June 24, 2009) (“The only thing that rises in falling markets is correlations.”); François Longin & Bruno Solnik, Extreme Correlation of International Equity Markets, 56 J. FIN. 649 (2001) (observing that correlation increases in bear markets, but not in bull markets, which implies that holding different asset classes acts may act as a drag on returns in bull markets without providing adequate diversification in bear markets). 124 See Whitehead, supra note 17, at 341 & n.85. 125 Id. at 347. 126 See id. at 346-52; see also Ian Ayres & Joshua Mitts, Anti-Herding Regulation, 5 HARV. BUS. L. REV. 1 (2015); Felix B. Chang, The Systemic Risk Paradox: Banks and Clearinghouses Under Regulation, 2014 COLUM. BUS. L. REV. 747. 127 See MARKUS BRUNNERMEIER ET AL., THE FUNDAMENTAL PRINCIPLES OF FINANCIAL REGULATION 8 (2009) (Geneva Reports on the World Economy, No. 11); Malcolm P. Baker & Jeffrey Wurgler, Do Strict Capital Requirements Raise the Cost of Capital? Bank Regulation, Capital Structure and the Low Risk Anomaly, 105 AM. ECON. REV. 315 (2015). 128 See Zeyu Zheng, Boris Podobnik, Ling Feng & Baowen Li, Changes in Cross-Correlations as an Indicator for Systemic Risk, 2 SCI. REPORTS 888 (2012). 129 See BRUNNERMEIER ET AL., supra note 127. 130 See Markus Brunnermeier & Tobias Adrian, CoVaR (2011) (Federal Reserve Bank of New York, Staff Report No. 348). 131 Viral V. Acharya, Christian Brownlees, Robert Engle, Farhang Farazmand & Matthew Richardson, Measuring Systemic Risk, in REGULATING WALL STREET: THE DODD-FRANK ACT AND THE NEW ARCHITECTURE OF GLOBAL FINANCE 87 (Viral V. Acharya, Thomas F. Cooley, Matthew P. Richardson & Ingo Walter eds., 2011); Viral V. Acharya, Lasse H. Pedersen, Thomas Phillippon & Matthew Richardson, Measuring Systemic Risk 1 (May 2010) (available at http://pages.stern.nyu.edu/~lpederse/papers/MeasuringSystemicRisk.pdf); Viral V. Acharya, Lasse Pedersen, Thomas Phillippon & Matthew Richardson, Taxing Systemic Risk, in REGULATING WALL STREET, supra, at 121. See generally Christian T. Brownlees & Robert F. Engle, SRISK: A Conditional Capital Shortfall Index for Systemic Risk Measurement (Jan. 1, 2015) (available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1611229); Robert F. Engle, Eric Jondeau & Michael Rockinger, Systemic Risk in Europe, REV. FIN. (forthcoming 2014) (available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2192536). See generally Viral V. Acharya, Robert F. Engle & Matthew P. Richardson, Capital Shortfall: A New Approach to Ranking and Regulating Systemic Risks, 102 AM. ECON. REV. 59 (2012). 132 See Aaron Brown, On Stressing the Right Size, GARP RISK REV., Nov./Dec. 2007, at 36.

Central Bank Journal of Law and Finance, No. 1/2015 41 The Promise and the Peril of Parametric Value-at-Risk (VaR) Analysis

133 Cf. WILLIAM BLAKE, THE MARRIAGE OF HEAVEN AND HELL 7 (1793) (“The road of excess leads to the palace of wisdom.”) (transcript available at https://en.wikisource.org/wiki/The_Marriage_of_Heaven_and_Hell). 134 One million divided by 252, the number of trading days in a year, is just under 4000. 135 Farber, Uncertainty, supra note 9, at 927. 136 See William D. Nordhaus, The Economics of Tail Events with an Application to Climate Change, 5 REV. ENVTL. ECON. & POL'Y 240, 242-43 (2011). 137 See Securities Act of 1933, Pub. L. 73-22, 48 Stat. 74 (enacted May 27, 1933, codified as amended at 15 U.S.C. §§ 77a-77aa); Securities and Exchange Act of 1934, Pub. L. No. 73-291, 48 Stat. 881 (enacted June 6, 1934; codified as amended at 15 U.S.C. §§ 78a-77pp). 138 See generally LINDA O. SMIDDY & LAWRENCE A. CUNNINGHAM, CORPORATIONS AND OTHER BUSINESS ORGANIZATIONS: CASES, MATERIALS, PROBLEMS 228-31 (7th ed. 2010) (describing the passage in the 1890s of enabling corporate statutes by New Jersey and Delaware). 139 See Xiong & Idzorek, supra note 51, at 29 (table 6) and 31 (figure 2) (reporting excess kurtosis of 9.50 in global high-yield bonds). 140 See Juan Salazar & Annick Lambert, Fama and MacBeth Revisited: A Critique, 1 AESTIMATIO 48, 64 (2010). 141 See Nordhaus, supra note 136, at 243. 142 Id. 143 Peter Conti-Brown, Commentary, A Proposed Fat-Tail Risk Metric: Disclosures, Derivatives, and the Measurement of Financial Risk, 87 WASH. U. L. REV. 1461, 1465 (2010). 144 Id. 145 See Nordhaus, supra note 136, at 253-54. 146 Id. at 256-57. 147 Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting).

42 Central Bank Journal of Law and Finance, No. 1/2015

Designing Macroprudential Mandates and Institutional Frameworks: The German Approach Andreas Guericke* Jochen Sprung**

Abstract

With effect from 1 January 2013 the German Financial Stability Committee was established. The Committee provides an institutional structure for cooperation between the Deutsche Bundesbank, the Federal Government, the Federal Financial Supervisory Authority (BaFin) and the Federal Agency for Financial Market Stabilisation in the field of financial stability. This article gives an overview of the Financial Stability Committee’s mandate, the leading role of the Deutsche Bundesbank and highlights the issues that should be considered when conferring macro-prudential powers to a central bank.

Keywords macro-prudential oversight; macro-prudential framework; German Financial Stability Committee; central bank

JEL Classification: G28, K23

* PhD, Head of the Deutsche Bundesbank’s Legal Department. ** Deutsche Bundesbank’s Legal Department, Senior Legal Counsel, responsible for legal issues relating to financial stability. The views expressed in this article are those of the authors and do not necessarily reflect the official stance of the Deutsche Bundesbank. It is based on a presentation held by Dr. Andreas Guericke on October 9, 2014 at the conference “Post Crisis Mandates for Central Banks: regulatory and institutional challenges” organized by the Central Bank of Romania in Bucharest.

Central Bank Journal of Law and Finance, No. 1/2015 43 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

1. INTRODUCTION

One of the main lessons learnt from the financial crisis was to strengthen the macro- prudential perspective when supervising the financial system1. Therefore, in the aftermath of the financial crisis, macro-prudential bodies were established. At EU level the European Systemic Risk Board (ESRB) took over its responsibilities in January 2011. The ESRB is part of the European System of Financial Supervision (ESFS) comprising the ESRB, the three European Supervisory Authorities (EBA, EIOPA, ESMA) and the competent national supervisory authorities. The ESFS was established at the proposal of the High Level Group on Financial Supervision in the EU published in February 2009 (“de Larosière report”)2. The ESRB is responsible for the macro-prudential oversight of the financial system within the EU. The ESRB instruments are, in particular, warnings and recommendations for remedial action, including for legislative actions. In reaction to legislative initiatives regarding macro-prudential frameworks in some Member States, the ESRB issued a recommendation “on the macro-prudential mandate of national authorities”3 at the end of 2011. The ESRB Recommendation aimed to provide guiding principles on core elements of national macro-prudential mandates, balancing the need for consistency among national approaches on the one hand and the flexibility to accommodate national specificities on the other hand4. A one-size-fits-all approach is inappropriate as the optimal design of national macro-prudential mandates depends on country-specific characteristics5. This ESRB Recommendation significantly influenced the design of the German macro- prudential framework which led to the establishment of the German Financial Stability Committee6. In June 2014 the ESRB published its assessment on the implementation of the ESRB Recommendation by the Member States on its website. The German approach has been assessed by the ESRB as fully compliant with the ESRB Recommendation. The German Financial Stability Committee as the macro-prudential authority set up in order to implement the ESRB Recommendation on the macro-prudential mandate of national authorities has to be distinguished from the so-called “designated authority” in the meaning of the CRR7 and CRD IV8. The macro-prudential authority in the sense of the ESRB Recommendation on the macro-prudential mandate of national authorities is not necessarily the same as the “designated authority” according to the CRR and the CRD IV9. The “designated authority” is responsible for executing the macro-prudential instruments laid down in the provisions of the CRR and the CRD IV. These comprise the possibility to impose stricter national measures (Article 458 CRR), the countercyclical buffer (Article 129 CRD IV), the additional buffer for global and other systemically relevant institutions (Article 131 CRD IV) as well as the systemic risk buffer (Article 133 CRD IV). In Germany, the Federal Financial Supervisory Authority (BaFin) is the designated authority in this sense.

44 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

2. THE GERMAN FINANCIAL STABILITY COMMITTEE

2.1. Institutional Design With effect from January 1, 2013 the German legislator established the German Financial Stability Committee10. The committee comprises representatives of four institutions. While the Deutsche Bundesbank, the Ministry of Finance and the Federal Financial Supervisory Authority (BaFin) nominate three representatives (each) who are endowed with voting rights, the Federal Agency for Financial Market Stabilisation participates with one representative in the committee without voting rights11.

2.2. Tasks The main tasks of the Financial Stability Committee are to deliberate on factors that are relevant for financial stability, to strengthen the cooperation between the institutions represented in the Financial Stability Committee in the case of a financial crisis, to advise on the handling of warnings and recommendations issued by the European Systemic Risk Board, to report annually to the lower house of Parliament, the Bundestag, and, if necessary, to issue warnings and recommendations addressing specific risks for financial stability in Germany12.

2.3. Powers In order to conduct its tasks, the Financial Stability Committee may draw attention to risks which might impair financial stability by issuing warnings and recommendations13. The latter includes measures that the Financial Stability Committee considers suitable and necessary for the addressee to implement in order to avert risks to financial stability14. The addressee of a warning or recommendation may be the Federal Government, BaFin or another public body in Germany15. As with ESRB’s recommendations, they are accompanied with a comply-or-explain mechanism which means that the addressee of a recommendation shall notify the Financial Stability Committee within a reasonable period of time of the manner how it intends to implement the recommendation. The addressee shall regularly inform the Committee of its implementation progress. If the addressee does not intend to implement a recommendation, it shall give detailed reasons on this matter16. In addition, the Financial Stability Committee has communication tools at its disposal. The Financial Stability Committee has the power to publish its warnings and recommendations17 as well as press releases18. Furthermore, the Financial Stability Committee shall report annually to the Bundestag on the situation and developments in financial stability as well as on its activities19. This annual report which serves first and foremost the accountability obligations related to the Parliament also provides the Financial Stability Committee with an opportunity to make the public aware of the risks for financial stability20.

Central Bank Journal of Law and Finance, No. 1/2015 45 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

2.4. Cooperation with the European Systemic Risk Board The Financial Stability Committee shall work closely with the ESRB and, insofar as it is necessary, with the authorities in charge of safeguarding financial stability in other Member States of the European Union21. In this respect, the Financial Stability Committee may exchange information with the ESRB and, insofar as it is necessary, with authorities charged with safeguarding financial stability in other Member States of the European Union insofar as this is required to safeguard financial stability22. As required by the ESRB Recommendation on the macro-prudential mandate of national authorities, the Financial Stability Committee shall inform the ESRB of its warnings and recommendations23. When warnings or recommendations may be expected to have a material cross-border impact, the Financial Stability Committee shall notify the ESRB prior to issuing its warning or recommendation24.

2.5. Role of the Deutsche Bundesbank Due to its expertise in macroeconomic analysis and financial markets, the Deutsche Bundesbank has a leading role and is tasked with important functions25. The Deutsche Bundesbank shall analyse factors that are relevant to financial stability as well as identify risks which may impair financial stability, propose warnings and recommendations, evaluate the implementation measures undertaken by the addressees in response to the recommendations and prepare the annual report of the Financial Stability Committee for the Parliament26.

This eminent role27 is recognised by conferring veto powers to the representatives of the Deutsche Bundesbank in the case of important decisions. Therefore, decisions concerning warnings and recommendations and the annual report may not be taken contrary to the votes of the Deutsche Bundesbank representatives28.

3. ROLE OF CENTRAL BANKS IN MACRO-PRUDENTIAL POLICY

At international level a common understanding has been developed, according to which central banks should play an important role with regard to macro-prudential policy29 due to their macroeconomic expertise and existing responsibilities in the area of financial stability. The Treaty on the Functioning of the European Union (TFEU) confers to the Eurosystem’s central banks the task to contribute to the smooth conduct of policies pursued by the competent authorities relating to the prudential supervision of credit institutions and stability of the financial system30. At national level, the Deutsche Bundesbank shall also contribute to safeguarding the stability of the financial system31.

46 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

However, the degree of a further involvement of central banks in macro-prudential policy as well as the type of instruments differs. In principle, two institutional models exist: The central bank model, where the responsibility for financial stability is transferred solely to the central bank32, and the committee model33 which stands for a common responsibility of several national institutions with different responsibilities in the field of financial stability, which should cooperate to ensure financial stability. The central bank model is common in countries where the central bank is in charge of micro-prudential supervision34. Only in a few cases the respective national supervisory authority - which is different from the central bank - was designated by the legislator as the macro-prudential authority in the sense of the ESRB recommendation on the macro-prudential mandate of national authorities35. In this case it would be difficult to ensure an important role of the central bank in macro- prudential policy. Regarding the macro-prudential authority’s powers, the legislators can choose to furnish the relevant national authority with legally binding and/or legally non-binding instruments.

3.1. The Central Bank as the Macro-Prudential Authority The macroeconomic expertise and the existing responsibilities of central banks in the field of financial stability might be crucial for the legislator to choose the central bank model. Furthermore, this model allows for information sharing and coordination between monetary and macro-prudential policy36. However, the conferral of the responsibility for financial stability solely on the central bank and, in particular, by vesting it with the competence to make use of legally binding instruments for macro-prudential purposes has to be weighed against the following considerations: The pursuit of more than one objective by the central bank might result in a conflict of interests37. Since, according to Article 127 paragraph 1 of the TFEU and Article 2 of the Statute of the ESCB and the ECB, the primary objective of ESCB central banks is to maintain price stability38. In case of a conflict of interests, the primary goal prevails over other central bank tasks, including macro-prudential policy. Such conflict of interests between monetary policy and macro-prudential policy might, for example, occur in case of a low interest rate environment. While the objective to maintain price stability may suggest an expansionary monetary policy, financial stability concerns may require a policy tightening in order to avoid the emergence of financial stability risks caused by price bubbles in asset markets39. Extending a central bank’s mandate beyond monetary policy, to macro-prudential policy as well, might call the broad independence, which central banks enjoy, into question40, in

Central Bank Journal of Law and Finance, No. 1/2015 47 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach particular if the central bank is empowered to take measures which are of fiscal relevance. Firstly, taking into account the principle of democratic control, conferring powers with effects for taxpayers’ money to an institution goes along with stronger accountability41, which might result in political bodies trying to influence the central bank42. Secondly, the experience with the financial crises and the state debt crisis demonstrates that broad competences of a central bank invite the political bodies to rely on the central bank instead of adopting out-of-favour political measures by themselves. Therefore, the central bank can come under pressure to act instead of political bodies, although the toolkit of the political bodies would be better suited to solve the problems. This might also result in increasing moral hazard43 if the financial market participants expect the central bank to act in any case of financial disturbances. When considering conferring macro-prudential powers to a central bank, it should also be taken into account that failures in macro-prudential policies might affect the central bank’s credibility with regard to monetary policy as well44. Even if a central bank, which is also the banking supervisory authority, is vested with comprehensive macro-prudential powers, the coordination with the government as a legislator is necessary45, insofar it has to be ensured that the central bank’s independence would not be affected. As many central banks have experience in the field of banking supervision, it seems likely that such central banks, in carrying out their macro-prudential tasks, concentrate on the banking sector rather than on the entire financial sector.

3.2. The Supervisory Authority as Macro-Prudential Authority One could ask why the overall responsibility for financial stability should not be transferred to a supervisory authority. In Germany, the supervisory authority (BaFin) is responsible for the supervision of the banking and insurance sector as well as of the financial markets. However, conferring the overall responsibility for financial stability to the supervisory authority would not sufficiently take into account that there are several players with responsibilities concerning macro-prudential issues, in particular the Deutsche Bundesbank’s expertise and experience in financial stability.

3.3. A Committee as the Macro-Prudential Authority The committee model ensures cooperation among those institutions whose actions have a material impact on financial stability46. This model takes into account the complex nature of financial stability, in particular the fact that measures in different policy areas can have an impact on financial stability47. This is the reason why the Financial Stability Committee has been given the power to issue warnings and recommendations, while the

48 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

entities involved in this committee have retained their competencies48. This concept ensures that the Deutsche Bundesbank’s independence is not affected in the execution of its tasks related to the Financial Stability Committee49. As mentioned above, the Financial Stability Committee consists of the representatives of the Deutsche Bundesbank, BaFin, the Federal Ministry of Finance and - in an advisory role - a representative of the Federal Agency for Financial Market Stabilisation (FMSA). The Financial Stability Committee, therefore, ensures the pooling of expertise from various institutions with responsibilities in financial supervision and financial market regulation50. The Deutsche Bundesbank is involved due to its macro-economic expertise and existing responsibilities in the area of financial stability. The BaFin participates as the responsible national supervisory authority for banks51, financial services providers, insurance undertakings and securities trading. As the ultimate responsibility for financial stability lies with the government, as taxpayer’s money is at risk52, it should be represented in the Financial Stability Committee. The participation of the Federal government (Finance Ministry) is also justified due to its right to initiate legislation which ensures that any necessary changes to the regulatory framework - in order to preserve financial stability - can be initiated in a timely manner53. Furthermore, macro-prudential policy interacts with fiscal policy; in particular, fiscal policy measures can have an impact on financial stability, e.g. tax policy decisions may lead to shifts in financial resources between different types of investment54. The Federal Agency for Financial Market Stabilisation (FMSA) contributes to the German Financial Committee with its expertise, especially in the recapitalisation and restructuring of banks. At first glance, one could question whether the involvement of the Federal Ministry of Finance bears the risk that decisions are taken on purely political reasons. In this regard, it is also discussed whether the involvement of many authorities results in a risk of inaction bias and compromises55. However, with its veto right concerning important decisions, the Deutsche Bundesbank has the power to avoid politically motivated decisions. The conferral of the competence to issue warnings and recommendations should not be interpreted as a weakness of the Financial Stability Committee, although the often used term “soft powers” might suggest it. The advantage of the power to recommend actions, coupled with a comply-or-explain-mechanism, is that it is broader than “hard instruments” 56. Recommendations can be used for the whole range of actions that can be taken by the addressees. However, it is important for the recommendation power to be coupled with a comply-or-explain-mechanism for the effectiveness of recommendations,

Central Bank Journal of Law and Finance, No. 1/2015 49 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

as it makes compliance more likely57. In this regard, the first ESRB recommendations seem to prove this assumption58. The committee structure has also the advantage that it ensures necessary information- sharing between the responsible entities.

4. CONCLUSION

Before conferring macro-prudential responsibilities on a central bank, the legislator should consider the above mentioned advantages and disadvantages, as well as the alternative models. The German legislator did not confer the financial stability function solely on the central bank, but chose the committee model with a leading role of the Deutsche Bundesbank. This model acknowledges, on the one hand, the Deutsche Bundesbank’s expertise and experiences in contributing to financial stability while respecting its independence and avoiding a conflict of interests with its primary objective of maintaining price stability. On the other hand, this model also takes the role of the government into account as it is ultimately responsible for financial stability.

50 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

NOTES

1 Report of the High Level Group on Financial Supervision in the EU (“de Larosière report”), February 2009, page 39, No. 153; Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report April 2013, page 41; Posch/van der Moolen, The macro-prudential mandate of national authorities, in: ESRB Macro-prudential commentaries, March 2012, page 1. 2 Report of the High Level Group on Financial Supervision in the EU (“de Larosière report”), February 2009. 3 Recommendation of the ESRB of 22 December 2011 (ESRB/2011/3). 4 Recital 4 of the Recommendation of the ESRB of 22 December 2011 (ESRB/2011/3). 5 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 86; Posch/van der Moolen, The macro-prudential mandate of national authorities, in: ESRB Macro-prudential commentaries, March 2012, page 6. 6 Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report April 2013, page 47. 7 Regulation (EU) No 575/2013 of the European Parliament and of the Council of 26 June 2013 on prudential requirements for credit institutions and investment firms and amending Regulation (EU) No. 648/2012. 8 Directive 2013/36/EU of the European Parliament and of the Council of 26 June 2013 on access to the activity of credit institutions and the prudential supervision of credit institutions and investment firms, amending Directive 2002/87/EC and repealing Directives 2006/48/EC and 2006/49/EC. 9 ESRB, Follow-up Report – overall assessment, June 2014, page 16, No. 6.3, table 11. 10 Act on Monitoring Financial Stability (“Financial Stability Law”) of 28 November 2012 (Federal Law Gazette I, page 2369) amended by Article 21 of the Act implementing Directive 2011/61/EU on Alternative Investment Fund Managers (Gesetz zur Umsetzung der Richtlinie 2011/61/EU über die Verwalter alternativer Investmentfonds (AIFM-Umsetzungsgesetz – AIFM-UmsG)) of 4 July 2013 (Federal Law Gazette I, page 1981). 11 Section 2 paragraph 3 Financial Stability Act. 12 Section 2 paragraph 2 Financial Stability Act. 13 Section 3 paragraph 1 and 2 Financial Stability Act. 14 Section 3 paragraph 2 Financial Stability Act. 15 Section 3 paragraph 3 Financial Stability Act. 16 Section 3 paragraph 4 Financial Stability Act. 17 Section 3 paragraph 6 Financial Stability Act.

Central Bank Journal of Law and Finance, No. 1/2015 51 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

18 Section 9 paragraph 4 of the Rules of Procedure of the Financial Stability Committee. 19 Section 2 paragraph 9 Financial Stability Act. 20 The first report of the Financial Stability Committee is available in German: http://www.bundesfinanzministerium.de/Content/DE/Downloads/Abt_7/2014-06-18-Bericht- AFS.pdf?__blob=publicationFile&v=4 21 Section 4 paragraph 1 Financial Stability Act. 22 Section 4 paragraph 2 Financial Stability Act. 23 Recital 9 and Recommendation B.4 of Recommendation of the ESRB of 22 December 2011 (ESRB/2011/3). 24 Section 4 paragraph 3 Financial Stability Act. 25 Explanatory Memorandum to the Act on Monitoring Financial Stability (Gesetz zur Überwachung der Finanzstabilität), Federal Law Gazette I (Bundesgesetzblatt I), 17/10040, page 12); see also Dombret, German lessons for Europe, Central Banking 24 (2013), 1, page 32. 26 Section 1 paragraph 1 Financial Stability Act. 27 The IMF came to the conclusion that the Deutsche Bundesbank has the leading role within the Financial Stability Committee (IMF, Key aspects of macroprudential policy, 10 June 2013, page 46, No 115). 28 Section 2 paragraph 5 Financial Stability Act. 29 IMF, Key aspects of macroprudential policy, 10 June 2013, page 29, No 81; Praet, The (Changing) Role of Central Banks in Financial Stability Policies, in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 4. According to recital 24 of Regulation No. 1092/2010 of the European Parliament and of the Council of 24 November 2010 on European Union macro- prudential oversight of the financial system and establishment of a European Systemic Risk Board, “the ECB and the national central banks should have a leading role in macro-prudential oversight”; Borio, (Too) great expectations, Central Banking 25 (2014), 1, page 84 states the opinion that “assigning a core role to central banks should be a priority”. 30 Article 127 paragraph 5 TFEU and Article 3.3 Statute of the ESCB and the ECB. 31 Section 1 paragraph 1 Financial Stability Act. 32 BE, CY, CZ, EE, GR, HU, IE, LV, LT, MT, PT, SK and UK chose or intend to choose the central bank model (ESRB, Follow-up Report – overall assessment, June 2014, page 14, No. 6.2, table 9). 33 AT, BG, HR, DK, FR, DE, IT, LU, NL, PL, RO, SI and ES chose or intend to choose the committee model and entrusted a board as the macro-prudential authority in the sense of the ESRB Recommendation on the macro-prudential mandate of national authorities (ESRB, Follow-up Report – overall assessment, June 2014, page 14, No. 6.2, table 9). 34 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 86.

52 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

35 Only FI and SE decided to designate their national supervisory authority as the macro-prudential authority in order to implement the ESRB Recommendation on the macro-prudential mandate of national authorities (ESRB, Follow-up Report – overall assessment, June 2014, page 14, No. 6.2, table 9). 36 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 86; Praet, The (Changing) Role of Central Banks in Financial Stability Policies, in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 4; Smets, Financial stability and monetary policy: How closely interlinked, International journal of central banking, Vol. 10 (2014), No. 2, page 287. 37 ESRB, Allocating macro-prudential powers, in: Reports of the Advisory Scientific Committee, November 2014, page 13; O’Sullivan/Kinsella, Navigating uncharted waters: central banks and financial stability, Financial regulation international, Issue 16.9 (2013), page 9; Visco, The challenges for central banks, Central banking 25 (2014), page 50. 38 ECB, Convergence Report, June 2014, page 21. 39 See ESRB, Allocating macro-prudential powers, in: Reports of the Advisory Scientific Committee, November 2014, page 7. 40 Buiter, in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 50; IMF, Monetary Policy in the New Normal, Staff Discussion Note, April 2014, page 23 and 27; see also Sibert, The Way Forward – Central Banks with Financial Stability Mandates: The Case of the Eurosystem, in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 438 and 444; 41 ESRB, Allocating macro-prudential powers, in: Reports of the Advisory Scientific Committee, November 2014, page 13 and 16; See also IMF, Monetary Policy in the New Normal, Staff Discussion Note, April 2014, page 24. 42 See Smets, Financial stability and monetary policy: How closely interlinked, International journal of central banking, Vol. 10 (2014), No. 2, page 277. 43 Visco, The challenges for central banks, Central banking 25 (2014), page 50. 44 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 86; Smets, Financial stability and monetary policy: How closely interlinked, International journal of central banking, Vol. 10 (2014), No. 2, page 287. In this regard see also Caruana, Redesigning the central bank for financial stability responsibilities, speech on the occasion of the 135th Anniversary Conference of the Bulgarian Bank, 6 June 2014, page 2. 45 Recommendation B.2 of Recommendation of the ESRB of 22 December 2011 (ESRB/2011/3). 46 Dombret, German lessons for Europe, Central Banking 24 (2013), 1, page 33. 47 Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report April 2013, page 47; Posch/van der Moolen, The macro-prudential mandate of national authorities, in: ESRB Macro-prudential commentaries, March 2012, page 3.

Central Bank Journal of Law and Finance, No. 1/2015 53 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

48 Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report April 2013, page 47; Dombret, German lessons for Europe, Central Banking 24 (2013), 1, page 33. 49 See also section 1 paragraph 2 Financial Stability Act. 50 Explanatory Memorandum to the Act on Monitoring Financial Stability (Gesetz zur Überwachung der Finanzstabilität), Federal Law Gazette I (Bundesgesetzblatt I), 17/10040, page 15). 51 As of 4 November 2014, competences of the BaFin regarding the supervision of credit institutions, in particular those concerning significant institutions, are transferred to the ECB by the Council Regulation (EU) No. 1024/2013 of 15 October 2013 conferring specific tasks on the European Central Bank concerning policies related to the prudential supervision of credit institutions. 52 Buiter, in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 42. See also Horáková/Carlyle/Ahlund, New mandates raise operational challenges. Central bankers face a raft of policy and operational challenges as their institutions take on new mandates and expanded responsibilities, Central banking, 24 (2013), 2, page 66. 53 Explanatory Memorandum to the Act on Monitoring Financial Stability (Gesetz zur Überwachung der Finanzstabilität), Federal Law Gazette I (Bundesgesetzblatt I), 17/10040, page 15); see also IMF, Key aspects of macroprudential policy, 10 June 2013, page 30, No. 86. 54 For example: The tax treatment of mortgage rates may incentivize investments in real estate which, in turn, may encourage excesses on the real estate markets (Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report April 2013, page 42). 55 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 87. 56 IMF, Key aspects of macroprudential policy, 10 June 2013, page 28, No. 75. 57 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 87. 58 Collin/Druant/Ferrari, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 88.

54 Central Bank Journal of Law and Finance, No. 1/2015 Andreas Guericke, Jochen Sprung

REFERENCES

1. Borio, Claudio E.V., (Too) great expectations? Macro-prudential policies are a welcome response to the financial crisis, but not necessarily in their incarnation. The author highlights the limitations of the tools available to central banks, in: Central Banking 25 (2014), 1, page 79 ff 2. Buiter, Willem H., The Role of Central Banks in Financial Sta-bility: How Has It Changed in: Evanhoff/Holthausen/Kaufmann/Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 11 ff. 3. Caruana, Jaime, Redesigning the central bank for financial stability responsibilities, speech on the occasion of the 135th Anniversary Conference of the Bulgarian Bank, 6 June 2014 4. Collin, Marianne/Druant, Martine/Ferrari, Stijn, Macroprudential policy in the banking sector: frameworks and instruments, in: Financial Stability Review, National Bank of Belgium, 2014, page 85 ff 5. De Larosière, Jacques (chair), Report of the High Level Group on Financial Supervision in the EU, chaired by Jacques de Larosière, Brussels, February 2009 6. Deutsche Bundesbank, Macroprudential oversight in Germany: framework, institutions and tools, Monthly Report, Frankfurt am Main, April 2013 7. Dombret, Andreas R., German lessons for Europe. Germany’s approach to macro- prudential oversight could offer insights into how to democratise “hard” policy tools and enshrine independence at a time when more power is being transferred to the ECB, in: Central Banking 24 (2013), Volume 1, page 31 ff 8. European Systemic Risk Board, ESRB Recommendation on the macro-prudential mandate of national authorities (ESRB/2011/3), Follow-up Report – Overall assessment, June 2014; Allocating macro-prudential powers, in: Reports of the Advisory Scientific Committee, November 2014 9. Horáková, Martina; Carlyle, Tristan; Ahlund, Arvid, New mandates raise operational challenges. Central bankers face a raft of policy and operational challenges as their institutions take on new mandates and expanded responsibilities, Central banking, 24 (2013), Volume 2 10. International Monetary Fund, Key aspects of Macroprudential Policy, June 2013; Monetary Policy in the New Normal, Staff Discussion Note, April 2014 11. O’Sullivan, Vincent; Kinsella, Stephen, Navigating uncharted waters: central banks and financial stability, Financial regulation international, Issue 16.9, November 2013 12. Posch, Michaela; Van der Molen, Remco, The macro-prudential mandate of national authorities, in: ESRB Macro-prudential commentaries, March 2012 13. Praet, Peter, The (Changing) Role of Central Banks in Financial Stability Policies, in: Evanhoff; Holthausen; Kaufmann; Kremer, The Role of Central Banks in Financial

Central Bank Journal of Law and Finance, No. 1/2015 55 Designing Macroprudential Mandates and Institutional Frameworks: The German Approach

Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 3ff 14. Sibert, Anne, The Way Forward – Central Banks with Financial Stability Mandates: The Case of the Eurosystem, in: Evanhoff; Holthausen; Kaufmann; Kremer, The Role of Central Banks in Financial Stability: How Has It Changed, in: World Scientific Studies in International Economics, 30, 2013, page 433 ff 15. Smets, Frank, Financial stability and monetary policy: How closely interlinked?, in: International journal of central banking, Vol. 10 (2014), No. 2, page 263 ff 16. Visco, Ignazio, The challenges for central banks. Demands are being made for central banks to consider financial stability alongside price stability, as a key component of their monetary policy. But that is nothing new for central banks, in: Central banking 25 (2014), Volume 1, page 49 ff

56 Central Bank Journal of Law and Finance, No. 1/2015

Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan)1

Lucian Croitoru*

Abstract

In this paper we show that monetary policies aiming at inflation targeting could not unwind global imbalances in an orderly manner if the GDP level is at potential and growth rates equal potential rates in deficit countries. Instead, as long as output is below potential, easing monetary policies could result in bringing global imbalances to pre-crisis levels. This outcome depends on the state of domestic demand in surplus countries. In the particular case of constrained domestic demand in countries with excess savings, monetary policy easing could result in currency depreciation and, thus, in higher exports and even larger excess savings.

Keywords global imbalances, exchange rate, monetary policy, excess savings

JEL Classification: E52, E58, F01, F21, F40

1 Published first in Romanian under the title Euro, dolarul, yenul și yuanul (The euro, the dollar, the yen and the yuan) * Monetary Policy Senior Advisor to the Governor of the National Bank of Romania

Central Bank Journal of Law and Finance, No. 1/2015 57 Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan)

1. INTRODUCTION

Nowadays some economists criticise the easing of monetary policies aimed at re- launching economic growth and bringing output back to potential. For instance, it seems to be widely accepted that the quantitative easing policies pursued by the Fed, the Bank of England, the Bank of Japan or the ECB, as well as the relaxation of the monetary policy in China will re-establish the global imbalances that played a part in the outburst of the crisis, but which were afterwards reduced by it.

The concern related to the monetary policies of the aforementioned central banks comes from the fact that the countries or the area they represent occupy a special place in the constellation of current account imbalances of the more than 220 world economies. Figure 1 shows that the economies with the greatest imbalances are the US, with a deficit of savings relative to investment, and Germany, China and Japan, with an excess of savings. The latter either “finance” the current account deficit of the US (namely they witness net capital outflows) or accumulate reserves1. Implicitly, the US either record net capital inflows or reduce their international reserves. In other words, these countries are counterparts.

Source: author`s calculations; UNCTADstad Source: author`s calculations; the World Bank; FRED

58 Central Bank Journal of Law and Finance, No. 1/2015 Lucian Croitoru

The imbalances among these countries stood at high levels or widened up to the outbreak of the crisis. After the crisis, these counterparts retained their roles from before and merely reduced the size of the imbalances, especially under the influence of cyclical factors (collapse of private demand, decline in oil price, deleveraging, and the drop in asset prices). More precisely, China and Japan have remained the great financiers and the Eurozone has joined them (Figure 2), thus complicating things even more. In these countries/areas, private income exceeds expenditure by far and the savings resulting from this imbalance are much higher than investment. If the concurrent monetary policy easing in surplus economies and the likely strengthening of the US monetary policy result in wider imbalances between these economies, then there is an asymmetry with the pre-crisis state of affairs. Before the crisis, monetary policies could not be used to help in the orderly unwinding of imbalances requested by the IMF in its reports as early as 2006-20072.

In this context, two questions are warranted: first, “why could not monetary policies unwind global imbalances in an orderly manner before the crisis did it abruptly?” And second, “why could monetary policy easing now, after the crisis, bring the imbalances back to the pre-crisis levels?’’ Such questions cannot be answered “in principle”. One can only speak “in principle” about monetary policies when discussing the principles underlying the monetary policy conduct. These principles are described, for instance, by Mishkin (2007)3. Otherwise, the effects of monetary policies pursued in order to achieve inflation and/or employment goals depend on the actual conditions specific to each country or monetary area. These conditions concern demand relative to potential output, the difference between the natural rate of interest and the effective rate, private demand in relation to potential private income, inflation expectations, the difference between actual inflation and the inflation target, debt stock, deleveraging, etc.

2. WHY MONETARY POLICIES COULD NOT UNWIND GLOBAL IMBALANCES DURING A BOOM

To answer the first question, we will refer to the actual conditions in the US, the country with the largest current account deficit in the world. From 2002 to 2007, the time span before the crisis, output was at potential and the inflation was equal to the implicit inflation target. These conditions show that monetary policy was exactly where theory

Central Bank Journal of Law and Finance, No. 1/2015 59 Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan) says it should be and supposedly there were no reasons for strengthening or easing the policy. Consequently, the reduction of the public budget deficit alone could have contributed to the narrowing of the current account deficit, in an attempt to prevent a crisis-induced reduction. The US Government’s budget cash deficit was 2.2 percent of GDP in 2006 and 2.5 percent of GDP in 2007, while the current account deficit stood at 5.8 percent of GDP and 5 percent of GDP respectively.

However, in the aforementioned conditions, a reduction of the budget deficit would have been difficult and would not have had effects on private demand, which was higher than private income. Narrowing the budget deficit would have pushed output below potential and, admittedly, would have diminished inflation expectations. The central bank, which targets inflation, would have reacted by lowering the short-term interest rate and hence the long-term ones in order to ensure the achievement of the inflation target. Consequently, the asset price would have increased and the dollar would have depreciated. The increased value of assets would have prompted households to earmark more funds for spending and companies for investment. The depreciation of the currency would have led to the increase in exports4.

These decisions of the private sector would have compensated the reduction in output determined by the initial fiscal contraction, but the goal – to narrow the current account deficit – would have been met. However, in order to reach this result, the fiscal authority should have wanted to reduce economic growth through the initial diminution of the budgetary deficit, which would have been politically inconvenient. In the end, US could not have reduced the current account deficit by using macroeconomic policies.

Unlike the US, an economy with a current account surplus could successfully use macroeconomic policies in order to reduce its surplus. Without personalizing, let us suppose this economy increases at a relatively high rate, equal to the potential one, and the current output level is equal to the potential output. A fiscal relaxation – which, in principle, is politically acceptable – would increase public demand, but would not entail an increase in output because the central bank would have to increase the interest rate in order to reach the inflation target. In this case, the value of assets would decrease, therefore private consumption and investment expenditures would decrease as well. With some help from the policy rate hike, the currency would appreciate, slowing down exports and fostering imports, thus narrowing the current account surplus at the end of the day.

If the world economy only consisted of the two economies, the wider budget deficit in the surplus economy would lead to a rise in exports in the economy with a current account

60 Central Bank Journal of Law and Finance, No. 1/2015 Lucian Croitoru deficit. Companies in the deficit economy, prompted by the weaker domestic currency, will want to export more, thereby meeting the increased demand for imports of the country reducing its surplus.

But if there are several surplus economies and only one with a sufficiently wide deficit, as is the actual case in the world economy, then the companies in other surplus economies will want to export more as well. In this situation, the companies in the country with a current account deficit will be able to export more in the country reducing its surplus (thus narrowing their country’s account deficit) only if they are more competitive than those based in other surplus countries. In conclusion, it is not certain whether the financed country narrows its current account deficit when a country reduces its current account surplus.

3. WHY MONETARY POLICY EASING COULD BRING THE GLOBAL IMBALANCES TO PRE-CRISIS LEVELS

To answer the second question (“why could monetary policy easing now, after the crisis, bring the imbalances back to the pre-crisis levels?”), it is necessary to identify the actual post-crisis environment.

Basically, there have been some sweeping changes. In most economies, the level of demand is much lower than the potential output level, despite widening budget deficits up to limits no longer financeable by the market. Moreover, in developed economies, private sector demand has collapsed and is still way below potential private incomes. Against this background, the central banks in developed countries have first reacted by lowering the monetary policy rate to virtually zero. Then they proceeded to the extreme relaxation of the monetary policy through quantitative easing rounds in order to bring inflation to the implicit or explicit target level. Out of the countries mentioned in Figure 2, China was the only one which did not need to resort to quantitative easing, but it also loosened the monetary policy very much in order to stimulate investments5.

Whether monetary policy easing in these countries leads or not to wider external imbalances in the future depends on each economy’s actual conditions regarding the two components of demand, i.e. domestic and external. As long as the actual conditions do not allow monetary policy easing to stimulate domestic demand in countries with excess savings (current account surplus), but rather only lead to the depreciation of the currency,

Central Bank Journal of Law and Finance, No. 1/2015 61 Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan) external imbalances will increase. Recently, in an article published in the Financial Times, Martin Wolf (2015) showed that this could be the case in the period ahead.

In a nutshell, Wolf (2015) shows that “it is hard to believe in a sustainable domestic spending boom”, even amid the ECB’s quantitative easing rounds, “given the large debt overhangs […], the absence of fiscal expansion and the fact that households and businesses […] are reluctant to spend”6. The ECB’s policy would bear fruit only if the weaker euro prompted a boom in exports which would increase the current account surplus7.

Wolf (2015) does not see a recovery of domestic demand in China or Japan either. In the case of the former, the credit-fuelled investment boom has become unsustainable. In the case of Japan, the level of public debt has soared, which renders it difficult to continue the policy of fostering demand by widening budget deficits. However, the depreciation of the yuan and of the yen respectively would lead to a rise in exports, which would increase the current account surpluses, but would help economic growth8. On the other hand, the counterparts for these larger surpluses in the Eurozone, China and Japan cannot be emerging or developing countries, which “are not creditworthy enough”9, but the US, the most capable of taking the risk of large capital inflows.

I share most of the arguments and rationale in Wolf’s article. Indeed, given the actual state of domestic demand in the biggest economies with current account surpluses, the monetary policies pursuing quantitative easing could lead to higher current account surpluses. However, given the precarious conditions of domestic demand, the increase in such surpluses could be the only chance of a rise in output.

This chance is all the greater as, unlike in the Eurozone, in the US and the UK, i.e. the largest counterparts financed (Figure 1), the GDP has exceeded the pre-crisis levels, nearing potential even more, and the quantitative easing programs have ended. Therefore, the USD and the GBP will not be subject to depreciation pressure10. On the contrary, the combination of loose monetary policies in the financing areas (Eurozone, China, Japan) and a tightened monetary policy in the US11 will contribute to the appreciation of the USD, which will reflect in the deterioration of the US current account. The widened current account deficit in the US will remain the main counterpart of the widened surpluses in the Eurozone, Japan and China, relatively similar to the pre-crisis period.

62 Central Bank Journal of Law and Finance, No. 1/2015 Lucian Croitoru

However, we cannot wrap up the analysis by concluding that monetary policy easing in the Eurozone, China and Japan will ultimately lead to an increase in exports and hence in GDP in those countries. This could be the happy ending of monetary policy easing and of imbalance widening provided that the imbalances stabilize at the levels recorded when current output would reach the potential level again.

This happy ending is, however, hard to anticipate, for two reasons: (i) it is not certain whether the current rounds of monetary policy easing, having an impact especially on the currency depreciation channel, will lead to a boom in exports, capable of raising output to potential; (ii) even so, while increased exports will lead to a faster GDP growth rate, the larger excess savings will keep the real interest rate low or reduce it even more, including in the financed countries (the US and other countries with a current account deficit). Thus, the increase rate of GDP will be substantially higher than the real interest rate in all counterparts, which is enough to generate a new credit boom that would fuel a new asset price bubble (Tirole, 1985).

However, a bubble can also emerge if the interest rate is higher than the GDP increase rate in certain conditions that make speculative growth generate larger resources for investment (Caballero, Farhi and Hammour, 2006)12. In this case, the impact of a bubble on economic growth can favour welfare if it occurs in the late stage of speculative growth, once growth has consolidated and the probability for the growth path to crash has diminished (Caballero, Farhi and Hammour, 2006).

The downtrend in economic growth rates that started in the ‘70s in developed countries has been going on to date. In the absence of structural reforms seeking to revive economic growth in the long run, the trend may persist. A new bubble would only interrupt this trend once again. Lawrence Summers (2013) thinks that, in the US, the aggregate demand in the absence of a bubble was very low in the pre-crisis period and that it was raised to potential due to the 2002-2007 bubble. Robert Gordon (2014) believes, however, that the potential level of GDP has decreased very much in time and therefore economic growth remains subdued.

One way or another, a new asset price bubble will push output to potential and/or the potential rate of growth to relatively high levels13. Thus, the bubble will coexist with relatively high economic growth and with low and stable inflation. So may be the case in other developed countries as well. If so, we will have to learn to see the benefits of an economic bubble as well, not just the disasters occurring when it bursts.

Central Bank Journal of Law and Finance, No. 1/2015 63 Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan)

On the other hand, if demand in the absence of a bubble is not way below potential, as seems to be the case in emerging economies, non-euro area EU Member States included, the bubble will act differently. We will see again relatively high rates of economic growth, pushed by the economic bubble beyond potential, and current account deteriorations14.

Nonetheless, there will be a while until then. In the meantime, we will have to wait and see if the depreciation of the euro, the yuan and the yen can generate an export boom that makes a success out of monetary policy easing in the Eurozone, China and Japan.

4. CONCLUSIONS

In order to attain the specific objectives, monetary policies will always take into account the actual conditions of the living systems called national economies or monetary areas. Ilya Prigogine and Isabelle Stengers (1984) showed that living systems have “dissipative structures” (which allow the exchange with the outside world), make a choice each time they are at a “crossroads”, re-establish “order through fluctuations” and they are almost always “far from balance”. This is actually the normal state in nature. We should not crave for a world in which all economies have current account deficits virtually equal to zero. In such a world, capitals would be strictly controlled and the efficiency would be very low. The necessary resources to maintain balances at zero or, on the contrary, to support exaggerated imbalances exceed by far the available resources. The resources at our disposal and our aspirations regarding the living standards make it so that we are sufficiently “far from balance” most of the time. The US, China, Japan, the UK and the Eurozone are the largest “dissipative structures” of the global system. Our analysis has shown that, nowadays, these economies are not sufficiently “far from balance” to produce enough jobs and increase general welfare15. Current actual conditions of private demand call for wider external imbalances in these economies in order to revive economic growth. The monetary policies pursued to increase inflation and employment will lead to the necessary expansion of imbalances in the US and the UK, as well as in the Eurozone, Japan and China. These evolutions should not be a matter of concern. On the contrary, given the excess savings in some economic areas, the re-widening of current account imbalances may be the only way to increase production and jobs, at least for a while. However, when the sober expectations that are still guiding our actions inevitably give way to euphoric ones, the imbalances will widen beyond the level required to take economic growth to potential. Nonetheless, as shown in this article, macroeconomic

64 Central Bank Journal of Law and Finance, No. 1/2015 Lucian Croitoru policies will not help prevent the excessive widening of imbalances, nor can they be used for the orderly narrowing of imbalances. It will then become apparent that the recent years’ financial overregulation will be useless as well.

Central Bank Journal of Law and Finance, No. 1/2015 65 Monetary Policy and the Global Imbalances (the Euro, the Dollar, the Yen and the Yuan)

NOTES

1 Basically, the economies with surpluses either have net capital outflows or accumulate reserves, while the economies with a current account deficit either have net capital inflows or cut their international reserves. In principle, at global level, the sum of these net capital inflows and outflows and of the changes in reserves should be equal to zero. However, it seems that, in practice, the global imbalance is not equal to zero, which made Krugman and Obstfeld speak about “the mystery of the missing surplus” (Krugman and Obstfeld, 2003). 2 Mainly the World Economic Outlook and the Global Financial Stability Report, issued in autumn and spring so that they can be discussed at the Annual Meetings of the Board of Governors of the IMF. 3 Briefly, these principles are: (i) the interest rate influences aggregate demand, which in turn determines short-term output fluctuations; (ii) there is a trade-off between inflation and unemployment in the short term, (iii) in the long run, there is no such trade-off, meaning that money is neutral; (iv) economic agents’ expectations crucially influence the inflation and respond to monetary policy; (v) credible assessments of monetary policies are rule-based; and, finally, (vi) financial frictions influence the business cycle. 4 The condition discussed in the text, namely that output is at potential and the GDP growth rate is equal to the potential rate, does not represent a particular case. Even if the output level and the growth rate were above potential, we would still end up to the case discussed in the text. Assuming that the divine coincidence holds, monetary policy will increase the interest rate up to the level at which current output would regain its potential level. From this perspective, the reasoning within the text is valid. In other words, the coordinated reduction of imbalances remains a problem if, once output has sustainably decreased to potential, the current account deficit is still relatively high. 5 China’s share of fixed capital investment in GDP increased during 2007-2012 by 7.1 percentage points, from 39 percent in 2007 to 46.1 percent in 2012. This growth is unsustainable and was promoted to compensate for the decrease of the export share in GDP by 5.9 percentage points, from 38.3 percent in 2007 to 27.7 percent in 2012. 6 Indeed, in the Eurozone, as in all developed countries, for that matter, the rate of time preference rate has dropped considerably in the aftermath of the crisis, probably running below the interest rate level. Consequently, households and companies overall have a propensity for saving, with savings becoming excessive not only in Germany and other countries of the “economic north” of the Eurozone, but in some countries of the “economic south” as well. From a current account deficit of 1.5 percent of GDP in 2008, the Eurozone ended up having a surplus of 2.4 percent of GDP in 2013. The deflation that emerged in December 2014 could last until the second part of 2016, increasing the real interest rate (which is above its natural level anyway), which exacerbates the debt burden and accelerates deleveraging. All these together limit substantially the effect of quantitative easing on the increase in euro area domestic demand. 7 I presented a similar view in January 2012, showing that a truly expansionary monetary policy of ECB, which would lead to euro depreciation, cannot be avoided (Croitoru, 2013). Quantitative easing could lead to further euro depreciation, thus stimulating exports outside the Eurozone. The increase in exports will add new jobs, allowing countries time to implement the necessary structural reforms on the labour market and on the goods and services market. Nevertheless, the current account surplus of the Eurozone will widen given the increased exports. Thus, the increase in the current account surplus will help GDP growth, which will, however, remain weak until structural reforms pay off.

66 Central Bank Journal of Law and Finance, No. 1/2015 Lucian Croitoru

8 In China and Japan, the excess of savings continues to be high, even if it has narrowed after the crisis (Figure 1). In both countries, the high volume of savings reflects a low time preference rate, which indicates a perpetual delay of expenses. Given the propensity for saving, loose monetary policies will have effects especially on the channel of yuan and yen depreciation respectively, particularly in relation to the US dollar. An increase in current account surpluses is to be expected in these countries as well, given the depreciation of their currencies. But, here as well, the increase in surpluses that worries some economists is necessary in order to enhance the increase in GDP. 9 In EU emerging economies, predictable conditions tend to make necessary the further easing of monetary policies, which shows that private demand continues to remain relatively weak. The cut in interest rates could weaken the respective currencies, which will rather continue to stimulate exports, temporarily diminishing their role as financed counterparts. The dollar will remain strong, to the extent to which the US remains the main such counterpart. 10 Any delay in the UK synchronising the interest rate growth cycles with those in the US will probably translate into depreciation pressures on the GBP versus the US dollar. 11 Probably starting with June 2016. 12 In the classical approach, Tirole’s included (1985), bubbles are a positive event in an economy with a structure that leads to over-saving and overinvestment, as is the case now, for instance, in post-crisis China. Within this framework, the bubble merely absorbs resources that, otherwise, would lead to over-accumulation issues. In the framework described by Caballero, Farhi and Hammour (2008), investment increases with the evolution of the bubble, because there are enough financing resources generated by speculative growth, to which adds the optimism on their sustainability. The conditions mentioned by Caballero, Farhi and Hammour (2008), which allow this “growth-funding feedback” between the stepped-up output and effective available funding (feedback that de facto generates speculative growth) are (i) technological progress concurrent with expansion, (ii) fiscal surpluses generated by rules, (iii) weak economic growth in the rest of the world, and (iv) the easing of financing conditions by the very feedback between growth acceleration and increased available funding. 13 An asset price bubble can coexist with a step-up in the growth rate of potential output and an interest rate cut. Thus, an economic bubble can coexist with high growth equal to potential, while inflation remains low and stable, as was the case in the US and in other developed countries in the run-up to the crisis that broke out in July 2007. The mechanism whereby this coexistence can be reached depends on a combination of positive shocks in productivity and in the time preference rate and is described in Croitoru (2015). As shown by Mendoza and Terrones (2008), the credit boom required to fuel a bubble is preceded by a positive productivity shock, which lends support to the explanation based on positive shocks. 14 No bubble occurred in Romania, although in the last three years the condition mentioned in the text has been met. The explanation probably lies with the fact that the discrepancy between the two rates was small and that, in the absence of some solid capital inflows, the GDP growth rate was relatively low, although higher than the real interest rate. 15 Once the economic growth accelerates in these countries, positive implications for the economic growth will occur in other countries as well.

Central Bank Journal of Law and Finance, No. 1/2015 67 Lucian Croitoru

REFERENCES

1. Caballero, Ricardo; Emmanuel, Farhi; Mohamad, L. Hammour (2006), “Speculative Growth: Hints from the U.S. Economy”, American Economic Review 96(4): 1159-1192 2. Croitoru, Lucian (2013), “The Eurozone: An Inconvenient Truth”, Romanian Journal of Economic Forecasting 2/2013, p. 200 3. Croitoru, Lucian (2015), “’Stagnarea seculară’, ‚’bulele salvatoare’ și creșterea economică în România” [“‘Secular Stagnation’, ‘Saving Bubbles’ and Economic Growth in Romania”] http://www.bnro.ro/Studii,-analize,-puncte-de-vedere-4009.aspx, p. 30 4. Gordon, Robert (2014), “The turtle’s progress: secular stagnation meets the headwinds”, in Teulings, Coen and Richard Baldwin (ed.) (2014), “Secular stagnation: facts, Causes and Cures”, A VoxEU.org Book, CEPR Press 5. Krugman, Paul; Obstfeld, Maurice (2003), “International Economics: Theory and Policy”, Addison Wesley, Sixth Edition, p. 314 6. Mendoza, Enrique G.; Marco, E. Terrones (2008), “An Anatomy of Credit Booms: Evidence from Macro Aggregates and Micro Data”, NBER, Working Paper 14049, http://www.nber.org/papers/w14049. 7. Mishkin, Frederic S. (2007), “Will Monetary Policy Become More of a Science?” Working Paper 13566, http://www.nber.org/papers/w13566, NBER (October) 8. Prigogine, Ilya; Isabelle Stengers (1984), “Order Out of Chaos: Man’s New Dialog with Nature”, Bantam Books 9. Summers, Lawrence (2013), “IMF Fourteenth Annual Research Conference in Honor of Stanley Fischer”, Washington, DC, November 8, http://larrysummers.com/imf- fourteenth-annual-research-conference-in-honor-of-stanley-fischer 10. Tirole, Jean (1985), “Asset Bubbles and Overlapping Generations”, Econometrica, Vol. 53, No. 6. (Nov. 1985), pp. 1499-1528 11. Wolf, Martin (2015), “Unbalanced hopes for the world economy: it is futile to ignore the reality that we have an integrated global system”, Financial Times, February 17

68 Central Bank Journal of Law and Finance, No. 1/2015

The Essential Characteristic of a Central Bank – Independence

Ianfred Silberstein*

Abstract

This paper approaches the importance and significance of the central bank’s independence, by highlighting this concept in the context of legal doctrine and of relatively recent opinions of some specialists from the central banks of various countries. Based on the opinions expressed during international conferences held in Romania by experts with legal training, regarding the role of such institutions reflected in the realities of the last years, the study moves on to the approach of the concept from doctrinaire perspective supported with arguments in relevant papers in the field, with reference to both foreign and Romanian doctrine. In this context, the necessary attention is paid to the analysis of legal texts of the normative act in force – Law No. 312/2004 – which emphasizes the support of the idea regarding the independence of the National Bank of Romania that the Romanian lawmaker consecrates in the assembly of this law. The underlined conclusions, based on the analysis, certify the beneficial nature of the central bank’s independence in Romania, an institution falling under the heading of the European System of Central Banks.

Keywords central bank; central bank’s independence; central bank’s objectives and instruments; monetary policy; price stability; Law on NBR Statute; European System of Central Banks

JEL Classification: E50, G28

* Ph.D, Adviser, National Bank of Romania, president of the Association of Legal Advisers of the Banking and Financial System, president of the European Association for the Banking and Financial Law - Romania

Central Bank Journal of Law and Finance, No. 1/2015 69 The Essential Characteristic of a Central Bank – Independence

1. INTRODUCTION

Contemporary reality demonstrates the role sui generis of the central bank institution, which is more and more present and its contribution to a state’s economy as well as to the evolution of the national banking system becomes more and more relevant. In turn, the aforementioned system stands out due to its remarkable contribution to the development of a national economy in one direction or another. It can also have significant effects on regional economies and even on world economy. The financial crisis triggered after 2007, first in the USA and then in Europe, has drawn the attention on a different topic in the economic field, especially regarding the evolution of the banking-financial system in a world of change in which concerns related to the present and the future of this system prove to be present nowadays, including in the perception of central banks’ bankers.

2. THE CENTRAL BANK IN THE VIEW OF SPECIALISTS FROM SUCH INSTITUTIONS

Browsing through the standpoints expressed in the last years by colleagues from central banks on general or particular current aspects, I consider these opinions worth mentioning. Analysing the actual situation from USA regarding the financial crisis and the role of New York central banks, Richard Charlton1, assistant to the Deputy President of New York FED, showed in June 2009 that “the current financial crisis has represented a major concern for our institution for almost two years”, claiming that “New York FED’s measures in the context of the current crisis did not exclusively consist in the support offered to a series of large institutions or more important from the systemic standpoint”. After analysing the crisis situation and the central bank’s role in this context, the author stated: “The crisis has also demonstrated the importance of efficiently managing liquidity. We oversee the liquidity of the most important organizations daily and we discuss the companies’ strategies, the market evolutions and the risk involved in their management… I underlined the effects on liquidity under the circumstances of a stressed market; the crisis has also demonstrated the importance of the effective management of liquidity”. “The crisis has also displayed the inadequacy of risk management in several financial institutions” says the author, who also underlines that “the crisis has also proven the importance of an effective and timely management of risks, which must truly be performed in the case of financial institutions”. Moreover, representatives of central banks from Europe expressed their views which should be taken into consideration. Thus, after surveying the concerns in Italy and the

70 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

regulatory documents adopted lately, Raffaele D’Ambrosio from the Bank of Italy emphasized that “in the case of a financial crisis, it is not easy to separate the monetary policy measures from the oversight policy measures”, showing by means of legal arguments that “the Bank of Italy checks whether the banks’ capital is inadequate, assesses their plans for stabilisation and consolidation, as well as the policies regarding the quotas adopted by credit institutions. At the same time, the Bank of Italy evaluates: the capital adequacy and the risk profile of banks, whether the requested financial instruments can be taken into account for the oversight capital and for the redemption of financial instruments in order to check if the financial situation and the solvability of banks are jeopardized.”2 When analysing the situation from Germany, prof. Rainer Kulms, from “Max Planck” Institute in Hamburg, starts from the fact that “the repercussions of the subprime American crisis reached Germany in July-August 2007. German state banks and some of the private banks had invested in the subprime American market and, consequently, they incurred enormous losses… public money was injected in some of these banks in order to save them from the imminent insolvency” and he shows that “the German management of the crisis focused on state banks and on solvability problems of private banks with a systemic importance”. The study shows that the central bank and the banking supervisory authority “published their common administrative approach in the Guidelines regarding the application of quality maintenance within continuous supervision through Deutsche Bundesbank”, in which the latter is to develop a risk profile for each financial-banking institution. On the other hand, in 2009, a study commented by the Federal Ministry of Finances evaluates both the corporate governance failures in banks and the deficiencies of the regulation approach in the case of banks. Professor Kulms concludes: “The current financial crisis clearly demonstrates that the conventional legal scheme of banking supervision, insolvency and the laws of commercial entities have been inadequate in determining a balance between the short- term needs and the long-term perspectives of the policy, between macroeconomic emergencies and microeconomic instruments. The post-October 2008 anti-crisis measures have introduced ad-hoc mechanisms of fast implementation of the conditions imposed under a help scheme administrated by public authorities. The German legislator refrained from adopting compulsory measures as broad in scope as the stabilisation options or as the rules regarding the administration of banks under banking law from 2009, England.”3 In October 2007, the former general counsellor of the International Monetary Fund, François Gianviti, performed an analysis worth paying attention and showed that: “there are different causes for financial crises. Most of them are generated by the risk taken excessively and by the lack of adequate systems of prudential control.” As far as financial

Central Bank Journal of Law and Finance, No. 1/2015 71 The Essential Characteristic of a Central Bank – Independence stability is concerned, the author claimed that: ”the role of central banks in promoting monetary stability is well-known… modern legislations of central banks are usually applied in order to set monetary stability as the key-objective of central banks” and that “broadly speaking, financial stability means the stability of the financial system or… it also means that the financial system’s components should be able to exert their key- functions within the system, in a satisfying manner”. Further on, the author claims that: “the role of a central bank in preventing or solving a financial crisis, depends to a great extent on the power and authority it is conferred by legislation”, but he concludes by saying that “the financial instruments available to central banks are more suitable to fight the effects of a financial crisis than to prevent it.”4 When analysing the role of a central bank, it is interesting to consider the ideas expressed by legal advisors from central banks headquartered in different states. Jiao Mei, head of research within the Legal Department of the People’s Bank of China, wrote in 2011: “There must be an authority of functional regulation in every financial market, which should be responsible for supervising the macroeconomic evolution and for the systemic risk, with a mandate to communicate early warning signals to the appropriate regulating authorities.”5 Referring to the way in which the Bank of Russia focused its efforts on sustaining the liquidity of the banking sector during 2008-2009, when the most urgent task was to overcome the effects of the global economic and financial crisis, Victoria Stepanenko, head of division within the Legal Department of the Central Bank of the Russian Federation, showed that among the most effective measures taken by this institution are worth mentioning: reducing the minimum reserve requirements, easier access of credit institutions to the Bank of Russia’s refinancing instruments, legislative amendments to improve the situation regarding banks’ liquidity. In addition, during the financial crisis, Bank of Russia adjusted its policy regarding interest rates several times and offered financial support to banks during merges and acquisitions. When overviewing the relevant legislation adopted at that time, the author came to acknowledge the contribution of the central bank to the improvement of the situation in the banking sector and to realise that the measures taken helped bring closer the legal regulation of relationships in the banking sector to the international standards.6 Comparing the objectives of central banks all over the world, Agus Santoso, deputy director of the Directorate for legal affairs in the Bank of Indonesia emphasized the fact that the crisis triggered debates on the role and responsibility of central banks, showing that: “The role of central banks as an anchor for price stability remains unquestionable. However, in post-crisis environment, central banks are expected to ensure stability in a broader sense, not just in terms of price stability. In particular, nowadays it is more or less universally acknowledged that central banks should have a formal mandate for

72 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

financial stability. Central banks have more and more frequently the task to supervise systemic risk in the context as the ultimate supplier of liquidity, so now central banks must also focus on the risks from the entire system and on obtaining the integrated opinion of the prudential supervision at micro individual level, as well as at macro level; in the case of Indonesia, the International Monetary Fund recommended to the Bank of Indonesia to hold a mandate for the financial system stability.”7 In a period in which the financial crisis demonstrated its vast consequences at national, regional and international level, people were more concerned with the role of central banks and more attention was drawn to these institutions.

3. THE DOCTRINE REGARDING THE INDEPENDENCE OF THE CENTRAL BANK

The necessity for the existence of such institution is determined by the state’s need to benefit from an institution that could ensure and maintain price stability in society. To this end, the central bank has been given prerogatives related to defining and implementing the monetary policy and the exchange rate, as well as to ensure the authorisation, regulation and prudential supervision of credit institutions. The central bank’s dependence on another authority would have had the effect of subordinating its functions to some interests, which would have diminished the objectivity of the activity that a central bank must carry out. Such an approach ends up in restricting or annihilating the expected effects of ensuring a monetary policy and an exchange rate appropriate to the economic development and to society’s interests. The action of the central bank cannot be subordinated to temporary interests of the government from a certain moment, which volens nolens, can have a subjective approach determined by the party’s interests or by problems occurred in the management of the economy, by unpredictable situations or by image necessities. What would the subordination of the central bank’s functions to such conditions mean and what would be the costs for society in the future if, at least temporarily, the central bank’s action was subordinated to the will of a state’s authority? The doctrinaire analysis and the practical approach of the thesis regarding the central bank’s independence concerned specialists in the field and sooner or later resulted in the governors’ understanding of society’s general goal, which ended up in regulations that settled down this principle to the legislations of the main countries. The concept of the central bank’s independence is well-argued both theoretically based on arguments related to the economic efficiency for the entire society and practically based on the results from social practice which demonstrated that a truly independent central bank represents a pledge for the evolution of a market economy.

Central Bank Journal of Law and Finance, No. 1/2015 73 The Essential Characteristic of a Central Bank – Independence

That is the reason why specialised literature is interested in analysing the subject of the central bank’s independence from different standpoints. Thus, Alan S. Binder8 shows that the independence of a central bank involves two things: first, the central bank has the freedom to decide on the way to reach its objectives; second, its decisions are strong enough against any governmental branch that could cancel them. When arguing his opinion, he explains that the statement according to which the bank can decide on its own does not mean that it can choose its objective by itself. On the contrary, he considers that, in a democracy, it seems appropriate for the public authorities to set the objectives and then to train the central bank to reach them. If it is to become independent, the central bank must demonstrate considerable discretion regarding the manner in which it uses its instruments in order to reach its legislative objectives. The institution does not need the authority to set its objectives. Such an authority would mean providing the central bank with an unjustifiably great power. Therefore, it is considered that the elected representatives of the people are the ones to make such decisions and that the bank should further on serve the public will. Consequently, the central bank must have instrumental independence, namely the independence of the instruments used in reaching the objective, not the independence regarding the objective. Showing that, in the case of Bundesbank and Federal Reserve Bank, the objectives of the monetary policy are set by legislative means, the author underlines that these are inaccurate enough to require a notable interpretation on behalf of the central bank. Alan S. Binder raises an issue he considers naive, but crucial: why does the central bank have to be independent? He thinks the essence of his answer is disarmingly simple: by its very nature, monetary policy requires a long-term horizon. This horizon is determined by the fact that the effects of monetary policy on production and inflation come in very late, so the decision-makers do not see the results of their action for a long time. The author considers that both politicians and mass-media along with their audience do not show too much sympathy or patience when it comes to a long-term horizon in the monetary policy fields. If politicians made monetary policy, this would be “day-to-day” type, due to the temptation to gain as soon as possible at the expense of the future, which leads to excessive inflation. Therefore, many governments wisely try to depoliticise monetary policy, placing it in the hands of non-elected technocrats with a vast experience in the field and isolated from political games. The result proved to be beneficial, at least in the case of industrialised countries. Based on the assumption that monetary policy implies notable technical experience, it has been concluded that central banks which have the highest degree of independence are

74 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

more credible fighters against inflation, which can accomplish a disinflation with lower social costs.

In Keith Bain’s opinion9, all the attempts to evaluate the central banks’ degrees of independence are subjective and open to criticism. In order for a central bank to be considered an institution with a high degree of independence, a normal request would be for the operational independence to be accompanied by an anti-inflationist objective that should be included in the country’s constitution.

By conducting a thorough analysis of this subject, Augusto Graziani10 shows that the general opinion of both experts and common people is strongly in favour of a completely institutional and de facto autonomy of the central bank in relation with the central government. This opinion is based on the motivation according to which an independent central bank will ensure price stability, a task that a bank subordinated to the government would be incapable of completing. The author finds the support of the same opinion in the Maastricht Treaty, which stipulates in Article 105 that price stability is the main task of the European System of Central Banks. The dominant opinion reveals the existence of some divergences between the tasks a central bank is reasonably pursuing and the tasks more frequently pursued by a government. A government may desire the full involvement of personnel in order to obtain a consensus from electors; a temporary growth in the levels of the economic activity may be pursued when political elections approach; public expenditure may be used in order to favour one influent political party or another. There seems to be a widely spread opinion according to which any government may be tolerant of inflation or even pursue inflationist policies deliberately, if set free to pursue its chosen tasks. In fact, as far as the pursued policy is concerned, theoretical models have established that this is the result of a game between a central bank, whose target-function is price stability and a government whose objective-function includes political consensus and the revenue redistribution, ultimately manifesting a tolerant attitude towards inflation.11 The idea that the government’s intervention is, by definition, an inflation factor was put forth for the benefit of a theoretical context. If one admits that spontaneous actions of market forces generate a full use of the workforce, the expenditure deficit is no longer necessary in order to reach a balance on the labour market. Under these circumstances, the expenditure deficit becomes, by definition, a factor of requested inflation, although price stability requires a balanced budget. In general, governments are in favour of higher expenditures and against the increase in taxes, so the only way in which a government can reach its targets is to finance expenditures, through debts or by increasing the money stock. Both instruments require the cooperation of the central bank to a different extent and under different circumstances. The easily drawn conclusion is that, if a central bank

Central Bank Journal of Law and Finance, No. 1/2015 75 The Essential Characteristic of a Central Bank – Independence is an independent institution, then it will be free to pursue its spontaneous policy of price stability while the government is incapable to interfere and to distort its action. Another factor specific to a market economy is the exchange rate policy that in many countries creates an obvious conflict between the central bank’s independence and the governments’ intention to keep the decisions regarding the control of the exchange rates for themselves. At the same time, there might be situations in which the governments would be in favour of the increase in interest rates, in order to attract foreign capital and to redress the balance of payments, while the central bank would tend to reduce it in order to stimulate the level of economic activity. When monetary union was adopted, national central banks gave up any role related to national exchange rates because there are no such rates between the states of the European Union and the exchange rates for other countries fall under the European Central Bank’s competence. In the doctrine of developed countries, with market economy, there have been many attempts to measure the central bank’s independence in order to establish the correlation that can exist between its independence and monetary stability. A composite index was proposed within the doctrine12, obtained from a number of institutional factors, such as: who appoints the central bank’s governor and board, how long is the mandate, does the law support the bank in the case of a conflict with the government, to what extent and under what circumstances can the bank extend credit facilities for the government, who sets the bank rate and other similar institutional arrangements. The author points out that, in practice, even when it has an actual formal independence, a central bank will inevitably be the subject in the country’s political climate and its decisions will rarely contradict the general political line of the government although the central bank uses its total freedom of action when making decisions. Augusto Graziani asserts that, basically, the Bank of Italy has always had a formal independence, being a stock company whose stockholders were mainly the savings banks. The author of the study concludes that the delimitation between formal and real independence is questionable, given the fact that on many occasions the real independence of the Bank of Italy was much lower than the institutional provision suggesting it. Specialised doctrine approaches, both generically and based on cases, the application of the central bank’s independence principle in regulations and practice of the countries with market economy, emphasizing the manner in which it is accomplished so as the beneficial effects of an independent central bank would produce results for all the components of a viable economy.

76 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

In his recent work “Monetary policy”13, professor Silviu Cerna shows that “the economic reason of the central bank’s independence is to allow those in charge with monetary policy to apply their policy without being constantly judged by the government based on the short-term results” and he concludes that: “an important condition that the statute of the central bank should meet is to create a structure of incentives that would ensure the placement of that particular institution’s governance in a long-term perspective.”

4. THE REFLECTION OF INDEPENDENCE IN LAW NO. 312/2004 REGARDING THE NATIONAL BANK OF ROMANIA STATUTE

As a matter of fact, in Romania, after 1990, legislation in the field has continuously sought to improve the qualification of the central bank and the emphasis laid on the significance of its independence from the other institutions and public authorities, which could not have been otherwise than beneficial in the context of its fundamental objective and its main attributions.

All necessities determined by the takeover of the acquis communautaire and the statement as concrete as possible of the central bank’s independence in the provisions of the law regarding the National Bank of Romania compelled the adoption of the law by Romanian Parliament on 28 June 2004. This law grants special attention to the legal statute of the institution, restating in Article 1 para. (1) that the National Bank of Romania is the central bank of the Romanian state and in para. (2) that it is an independent public institution. The last collocation defines, on the one hand, the nature of public institution resulted from the fact that its entire capital belongs to the state [Article 38 para. (1)], as well as the essential function granted by the lawmaker, set as the fundamental objective, to ensure and maintain price stability, a general type of function in the interest of the entire Romanian society. In order to emphasise its specificity and the necessity to establish an independent statute from state authorities, the lawmaker has expressly consented the character that distinguishes the central bank from other public institutions. All provisions from the regulatory document ensure the necessary legal framework so that the National Bank of Romania should have the same legal regime as the one provided by other legislations of the countries with market economies. The idea of independence ensures to the central bank a distinct position among the public institutions. In fact, it may also be noticed within the Constitution of Romania, which consecrates in its provisions the legislative, executive and judiciary authorities,

Central Bank Journal of Law and Finance, No. 1/2015 77 The Essential Characteristic of a Central Bank – Independence whose separation is clearly defined, that the National Bank of Romania is not mentioned in the fundamental Law. Thus, the lawmaker creates a special law in order to regulate exhaustively all the aspects specific to this institution and to express the understanding of the role and fundamental objective specific to such an institution. The Law regarding the National Bank of Romania Statute reflects the lawmaker’s concern to enshrine by means of clear legal provisions the application of this principle regarding the central bank of Romania. Hence, from the very beginning, Article 1 defines it as an independent public institution, which reflects into the entire economy of the law. Article 2 para. (3) relates the obligation of the central bank regarding the support granted to the state’s general economic policy to the non-prejudice to its fundamental objective, which implies its priority regarding the obligation of support, thus this law reflects the independence of the central bank from this perspective as well. Article 3 para. (1) imperatively expresses the interdiction for the National Bank of Romania and its executive members to request or receive instructions from public authorities, which strengthens the idea of this institution’s independence. Article 6 para. (1) which “forbids the National Bank of Romania to purchase claims on the state, local and central public authorities from the primary market” and Article 7 para. (2) which “forbids the National bank of Romania to lend overdrafts or any other type of credit to the state, local and central public authorities…” support the independent nature of the central bank from other public authorities. Chapter VI of the law – Articles 27 – 29 expressly provides that the National Bank of Romania and the Ministry of Public Finances conclude agreements regarding the general current account of the State Treasury and operations with public securities, which demonstrates the equality of the parties in such legal rapports, thus underlining the fact the National Bank of Romania negotiates independently in all these cases. Chapter VIII - Management and administration – which shows that the members of the National Bank of Romania Board are appointed by the Parliament on the recommendation of the competent standing committees of the two Chambers of Parliament, highlights the fact that the Lawmaker has a decisive role in appointing and revoking the National Bank of Romania Board, without mentioning the Executive that cannot pronounce itself in this regard. Article 35 para. (4), which shows that the Governor submits to the Parliament the National Bank of Romania’s annual report, subject to debate, but not voted in the joint session of the two Chambers of Parliament, proves that the Parliament does not have the

78 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

authority to meddle in the central bank’s activity and it is only called to be informed about the contents of the documents presented by the National Bank of Romania. The provisions of Article 48 para. (1) are in the same line: “the National Bank of Romania is mandated to issue the necessary regulations to implement and to impose the compliance with the legal provisions.” Thus, the central bank has full freedom, determined by its experience and professional capacity, to issue the regulations it considers necessary in the field. As one may notice, many of the provisions of the current law on the Statute of the National Bank of Romania contain dispositions that support this characteristic of the central bank which entitles it to be treated as an institution sui generis of a public nature, but not subordinated to any other governmental authorities or institutions, which could interfere in the policy and activity of the National Bank of Romania.

5. THE OPINIONS OF THE EUROPEAN CENTRAL BANK REGARDING THE INDEPENDENCE OF THE CENTRAL BANK

The importance of the central bank’s independence is also emphasized in the European Central Bank’s various opinions expressed while analysing some draft legislative provisions of the Member States of the European Union, which argued, from this point of view, the position of the European Central Bank on these particular regulations. The European Central Bank (ECB) must be consulted by national authorities on any proposed national legislation that falls within its fields of competence, pursuant to Article 127 para. (4) and Article 282 para. (5) of the Treaty on the Functioning of the European Union (the ‘Treaty’) and Article 4 of the Statute of the European System of Central Banks and of the European Central Bank (the ‘Statute of the ESCB’). The scope of the obligation on national authorities to consult the ECB is set out in Council Decision 98/415/EC of 29 June 1998 on the consultation of the European Central Bank by national authorities regarding draft legislative provision. The obligation relates to any draft legislative provisions within the ECB’s fields of competence and Article 2 para. (1) of Council Decision 98/415/EC lists six main fields of competence. The legal opinions bear the signature of the European Central Bank’s President; therefore, the institution is held responsible for the content of each opinion stated.

Central Bank Journal of Law and Finance, No. 1/2015 79 The Essential Characteristic of a Central Bank – Independence

ECB opinions are relevant for the Member States when they are presented to the legislative authority along with the legislative proposal in order to be adopted and become a law. In this context, we think that some of the European Central Bank’s opinions stated in such cases in order to stress the nature of the central bank’s independence are worth mentioning. Thus, according to the European Central Bank’s opinion of 25 July 2006 (CON/2006/38), delivered at the request of the Bank of Greece with regard to a draft provision on its powers in the field of consumer protection, in terms of specific observations, with reference to the financial independence of the Bank of Greece, the ECB observed that the new tasks described in some paragraphs of the draft require the Bank of Greece to commit considerable human and financial resources, in order to take responsibility for other functions than those provided in the Statute. This determined the European Central Bank to hold the view that “Member States may not put their central banks in the position of not having sufficient financial resources to carry out their ESCB and Eurosystem related tasks, as applicable.” By planning the transfer of the additional powers, the European Central Bank expects the central bank to take over the responsibility to ensure that the operational capacity of the bank to carry out the tasks related to the Eurosystem would not to be affected. The ECB reiterated the same aspects in its opinion of 21 March 2007, delivered at the request of the Czech Republic’s Ministry of Industry and Trade on the draft law amending the Law on consumer protection relating to the Czech National Bank, concluding that the Czech authorities will need to ensure that the fulfilment of the additional tasks will not affect the operational capacity of the Czech National Bank to carry out ESCB-related tasks. The same considerations are also found in the European Central Bank’s opinion of 5 October 2007 (CON/2007/29), delivered at the request of the Portuguese Ministry of Finances, regarding the decree-law amending the legal framework on credit institutions and financial companies. The European Central Bank’s opinion of 21 January 2015 on the role of the National Bank of Slovakia in the resolution in the financial market (CON/2015/3) refers to the role of the central bank from this country and stating that, according to the principle of financial independence, a national central bank must have enough resources to perform both its

80 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

tasks related to the European System of Central Banks and its national tasks, such as the financing of its administrations and its own operations. Financial independence also implies that a national central bank should always be capitalized enough and should hold enough appropriate resources, both human and financial, so that such tasks would not affect the national central bank’s operational capacity related to its specific tasks. Special attention should be paid to the protection of the personal independence of the members of the national central bank bodies in terms of involving its representatives in decision-making supervisory bodies, which also stands for the case discussed. To the same end, the European Central Bank’s opinion of 2 March 2015 on the independence of the Bank of Slovenia and the members of the decision-making bodies (CON/2015/8) refers to the protection of the central bank’s independence with reference to Article 130 of the Treaty which establishes the principle of personal independence, an essential aspect of the central bank’s independence principle for the members of the European System of Central Banks. The European Central Bank’s opinion of 10 March 2015 on the legal framework of the National Bank of Poland (CON/2015/9), delivered at the request of the Ministry of Finances from this country, emphasizes that the European Central Bank’s Convergence Report 2014 identified some provisions from other laws that are incompatible with the requirements of the Treaty on the Functioning of the European Union with respect to central bank personal independence, in particular regarding the president of the bank and its members of the decision-making bodies. The European Central Bank’s opinion of 24 March 2015 on the credit mediator legal framework (CON/2015/12), delivered at the request of the Portuguese Ministry of Finances, on a draft law amending Decree Law No. 144/2009 of 17 June 2009 establishing the Portuguese credit mediator legal framework, shows that the European Central Bank encourages the Portuguese authorities to be aware of the resources implications of the draft law, as under the legal framework in force, which was not modified in this respect by the draft law, the resources will come from the Bank of Portugal. Although it is independent from the central bank, the latter pays for the services and provides the necessary technical, administrative and financial support necessary for the credit mediator and council, for the fulfilment of their duties. The financing and the support of the Bank of Portugal for the mediator is neither based on the Treaty or the Statute of the European System of Central Banks and of the European Central Bank, nor of a supervisory nature. Thus, the resources of the Bank of Portugal

Central Bank Journal of Law and Finance, No. 1/2015 81 The Essential Characteristic of a Central Bank – Independence could simultaneously be used for the mediation of credit relations between customers and credit institutions, as well as for the supervision of these institutions. This is the Bank of Portugal’s task, as the national supervisory authority, which is considered an amplified contradiction in terms of the restructuring of the financial debt affecting the credit institutions under the supervision of the Bank of Portugal. In its turn, the European Central Bank’s opinion of 28 May 2005 on the legal framework for the deposit guarantee scheme and resolution in the financial markets (CON/2015/17), delivered at the request of the Ministry of Finances from Poland, shows in relation to Poland’s obligations to implement Directive 2014/59/EU and Directive 2014/49/EU into the national law that these provisions can impact the role and the tasks of the National Bank of Poland as a central bank and as a member of the European System of Central Banks. The European Central Bank shows that the national legislation providing for the financing of a deposit guarantee scheme for credit institutions by a national central bank within the European System of Central Banks will only be compatible with the monetary financing prohibition, if such financing is short-term, it refers to emergency situations, systemic stability problems and decisions are made by the national central bank. The support of the national central bank for the deposit guarantee schemes could not consist in a systemic operation of initiating funds.

The European Central Bank’s opinion of 1 July 2015 on recovery and resolution in the financial market (CON/2015/22), delivered at the request of the Czech Republic’s Ministry of Finances, shows that the European Central Bank restates that resolution tasks discharged by central banks are not considered central banking tasks resulted from its independence, in accordance with Article 130 of the Treaty. At the same time, it also shows that the financial independence principle requires a national central bank to have enough financial resources, both operationally and financially, for the fulfilment of the tasks resulted from its quality of a member of the European System of the Central Banks and as far as its national tasks are concerned. From this standpoint, the European Central Bank appreciates the provisions from the draft law regarding the financing of some new tasks concerning the resolution attributed to the Czech National Bank. The European Central Bank states that it is important for the Czech National Bank’s Board, as responsible body of decision-making, to have the ultimate control over any decisions in the resolution field that can affect the financial independence of the Czech National Bank. The European Central Bank states that it is important the Czech National Bank’s Board as responsible body of decision-making can have ultimate control over any decisions in the resolution field that can affect financial independence of the Czech National Bank.

82 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

All these opinions of the European Central Bank demonstrate the European institution’s preoccupation with the observance of the independence of the national central banks belonging to the European System of Central Banks. The particular references in the field, contained in the Convergence Reports of the European Central Bank from June 2013 and June 2014, are also worth mentioning. Thus, the first report shows that the evaluation of the national legislation of the Member States of the European Union, especially of the statutes of the national central banks, has taken into account the concept of independence which includes financial, personal, institutional and functional independence that should be analysed separately. Their analysis has also been conducted by means of the policies adopted by the European Central Bank. Thus, the Convergence Report from June 2013 shows that “central bank independence is not an end in itself, but is instrumental in achieving an objective that should be clearly defined and should prevail over any other objective. Functional independence requires each national central bank’s primary objective to be stated in a clear and legally certain way and to be fully in line with the primary objective of price stability established by the Treaty. It is served by providing the national central banks with the necessary means and instruments for achieving this objective independently of any other authority… The principle of institutional independence is expressly referred to in Article 130 of the Treaty and Article 7 of the Statute... These prohibit the EU institutions, bodies, offices or agencies and governments of Member States any attempt of influencing those members of decision-making bodies of the national central banks whose decisions can influence the fulfilment of the Central National Bank’s tasks.” As far as personal independence is concerned, the report mentions Article 14.2 of the Statute which provides for the protection against the arbitrary dismissal of Governors by stipulating that “a Governor may be relieved from office only if he no longer fulfils the conditions required for the performance of his duties or if he has been guilty of serious misconduct”, with the possibility of recourse to the Court of Justice of the European Union. This provision is meant to further safeguard the independence of the central bank, given the fact that national central banks’ Governors are members of the General Council of the European Central Bank and will be members of the Governing Council upon the adoption of the euro by the Member States they belong to. At the same time, the report shows that amending the national legislation on the remuneration the members of the national central banks’ decision-making bodies may not be used as a way of seeking to influence the appointed members of the NCBs’ decision-making bodies. As a matter of principle, the legislative measures “should apply only for future appointments.”

Central Bank Journal of Law and Finance, No. 1/2015 83 The Essential Characteristic of a Central Bank – Independence

The document also shows that the national central bank’s independence would be endangered if it could not autonomously avail itself of sufficient financial resources to fulfil its mandate, even if from a functional, institutional and personal point of view it is fully independent. In this regard, it is important to note that the principle of financial independence requires a national central bank to have sufficient means to perform both its European System of Central Banks-related tasks and its national tasks. In this respect, the report concludes: “The concept of financial independence should be assessed from the perspective of whether any third party is able to exercise either direct or indirect influence not only over a national central bank’s tasks but also over its ability to fulfil its mandate, both operationally in terms of manpower, and financially in terms of appropriate financial resources”. These considerations are taken into account each time an analysis of the legislation on the national central bank from each Member State is performed. In relation to the legislation, the degree of independence of each central bank is estimated. If these requirements are not met, the European Central Bank notes it annually in its convergence report.

6. CONCLUSIONS REGARDING THE SIGNIFICANCE OF THE CENTRAL BANK’S INDEPENDENCE

This institution, as part of the European System of Central Banks, carries out its activity in compliance with the EU relevant legal framework, adapting its policy to the one decided at the ESCB level, in which all the central banks from the Member States of the European Union are integrated. Thus, the National Bank of Romania directs its policy towards the letter and the spirit of the European regulations sooner than governmental institutions, thus contributing actively to the integration of Romania in the European Union, to which it accessed on 1 January 2007.

In this context, the significance of its independence gains new meanings, contributing to the strengthening of the confidence transmitted by this institution nationally and abroad.

The central bank’s independence means that its policy, in relation to its fundamental objective and main attributions and in compliance with the legal regulations in the field, can determine an evolution of the national economy beyond temporary political interests, which most of the times leads it towards directions and modalities that do not maintain their viability in time.

84 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

The central bank’s independence proves to be much more useful to an Executive Power that understands how to respect it and to cooperate for the benefit of the society so that, from an objective perspective, the well-intended collaboration of the two institutions could create the necessary framework for the settling of the national economy on the right path for progress, fact proved by the examples in which their relationship was based on such coordinates.

In this context, the National Bank of Romania has shown its capacity of determining the reduction of the inflation up to zero, of ensuring the stability and viability of the banking system, including in a period of global financial-banking crisis at the world level.

Central Bank Journal of Law and Finance, No. 1/2015 85 The Essential Characteristic of a Central Bank – Independence

NOTES

1 Richard Charlton – “The Response of the Federal Reserve Bank of New York to the Current Crisis” in “Financial crisis across the world. The regulation and supervision of credit institutions”, Wolters Kluwer-Romania Publishing House, Bucharest 2010, pp. 7-15 2 See Raffaele D’Ambrosio – Aspects of the recent rules in Italy on the financial crisis in the light of the general principles of the Italian banking Law in the work cited above, pp. 16-25 3 See the same paper Rainer Kulms – The reaction of the German regulation entity, pp. 39-69 4 François Gianviti – The role of central banks in financial stability in Current problems in banking law, Wolters Kuwer Publishing House – Bucharest 2008, pp. 41-53 5 Jiao Mei – The legal framework for the reduction of risks in the banking sector from China in “Legal aspects in a changing banking system”, idem p. 111 6 Victoria Stepanenko – “The current legal regulation of banking activity” in op. cit. pp. 40-77 7 Agus Santoso – “The necessity for a mandate of the Central Bank for the treatment of financial stability” in op. cit. pp. 160-171 8 A. S. Binder – Central Banking in Theory and Practice, The MITPRESS, Cambridge, Massachusetts, London, England, 1988, p. 53 and the following 9 K. Bain – Some Problems with the Use of “Credibility” and “Reputation” to Support the Independence of Central Banks in the Political Economy of Central Banking, edited by Ph. Arestis and M. C. Sawyer, E. Elgar, Cheltenham, UK, Northampton MA, USA, 1988, p. 41 and the following. 10 A. Graziani – The Independence of Central Banks: the Case of Italy in the Political Economy of Central Banking, edited by Ph. Arestis and M. C. Sawyer, E. Elgar, Cheltenham, UK, Northampton MA, USA, 1988, p. 169 and the following 11 In this sense, the author cites: Gpiga & L. Pecci – Indexation, Inflation and Central Bank Independence, University of Rome, Department of Public Economics, Working Paper, 1997 12 Citing V. Grilli, D. Masciandaro & G. Tabellini – Political and Monetary Institutions and Public Policies in Industrial Countries, Economic Policy, 6(2) October, 1991, pp. 341-392 13 Issued by the Romanian Academy’s Publishing House, Bucharest 2014, p. 260 and the following

86 Central Bank Journal of Law and Finance, No. 1/2015 Ianfred Silberstein

REFERENCES

1. Bain, K., Some Problems with the Use of ʺCredibilityʺ and ʺReputationʺ to Support the Independence of Central Banks in the Political Economy of Central Banking, edited by Ph. Arestis and M.C. Sawyer, E. Elgar, Cheltenham, UK, Northampton MA, USA, 1988 2. Blinder, A.S., Central Banking in Theory and Practice, The MITPRESS, Cambridge, Massachusetts, London, England, 1998 3. Cerna, Silviu, Politica monetară, Ed. Academiei Române, București, 2014; [Monetary Policy, Romanian Academy’s Publishing House, Bucharest, 2014] 4. Charlton, Richard, Răspunsul lui Federal Reserve Bank of New York în fața crizei prezente în lucrarea ,,Criza financiară în lume. Reglementarea și supravegherea instituțiilor de credit, Ed.Wolters Kluwer-România, București, 2010; [The Response of the Federal Reserve Bank of New York to the Current Crisis in “Financial crisis across the world. The regulation and supervision of credit institutions”, Wolters Kluwer- Romania Publishing House, Bucharest, 2010] 5. D'Ambrosio, Raffaele, Aspecte ale recentelor reguli din Italia privind criza financiară în lumina principiilor generale ale Legii bancare italiene în lucrarea Criza financiară în lume. Reglementarea și supravegherea instituțiilor de credit, Ed. Wolters Kluwer- România, București, 2010; [Aspects of the recent rules in Italy on the financial crisis in the light of the general principles of the Italian banking Law in “Financial crisis across the world. The regulation and supervision of credit institutions”, Wolters Kluwer- Romania Publishing House, Bucharest, 2010] 6. Gianviti, François, Rolul băncilor centrale în stabilitatea financiară în Probleme actuale în dreptul bancar, Ed. Wolters Kluwer – România, București, 2008; [The Role of Central Banks in Financial Stability in Current Problems in Banking Law, Wolters Kluwer – Romania Publishing House, Bucharest, 2008] 7. Graziani, A., The Independence of Central Banks: the Case of Italy in The Political Economy of Central Banking, edited by Ph. Arestis and M.C. Sawyer, E. Elgar, Cheltenham, UK, Northampton MA, USA, 1988 8. Grilli, V. ; Masciandaro, D.; Tabellini, G., Political and Monetary Institutions and Public Policies in Industrial Countries, Economic Policy, 6(2) October, 1991 9. Kulms, Rainer, Reacția organismului de reglementare german în lucrarea ,,Criza financiară în lume. Reglementarea și supravegherea instituțiilor de credit”, Ed.Wolters Kluwer-România, București, 2010; [The reaction of the German regulation entity in “Financial crisis across the world. The regulation and supervision of credit institutions”, Wolters Kluwer-Romania Publishing House, Bucharest, 2010] 10. Mei, Jiao, Cadrul juridic pentru reducerea riscurilor în sectorul bancar din China în lucrarea Aspecte juridice într-un sistem bancar în schimbare, Ed.Wolters Kluwer – România, București, 2011; [The legal framework for the reduction of risks in the banking sector from China in “Legal aspects in a changing banking system”, Wolters Kluwer – Romania, Bucharest, 2011] 11. Santoso, Agus, Necesitatea unui mandat al Băncii Centrale pentru tratarea stabilității financiare, în lucrarea Aspecte juridice într-un sistem bancar în schimbare, Ed.Wolters

Central Bank Journal of Law and Finance, No. 1/2015 87 The Essential Characteristic of a Central Bank – Independence

Kluwer – România, București 2011; [The necessity for a mandate of the Central Bank for the treatment of financial stability in “Legal aspects in a changing banking system”, Wolters Kluwer – Romania, Bucharest, 2011] 12. Silberstein, Ianfred, Banca Națională a României – De la organ al administrației centrale la instituție publică independentă, Ed. Hamangiu București 2006; [National Bank of Romania – From a central administration entity to an independent public institution, Hamangiu Publishing House, Bucharest, 2006;] 13. Silberstein, Ianfred, Studii de drept bancar – Evoluții ale legislației bancare între anii 1995 – 2010, Ed. Wolters Kluwer România, București, 2010; [Studies of banking law – Evolutions of the banking legislation in 1995-2010, Wolters Kluwer Romania Publishing House, Bucharest, 2010;] 14. Stepanenko, Victoria, Prezenta reglementare juridică a activității bancare, în lucrarea Aspecte juridice într-un sistem bancar în schimbare, Ed.Wolters Kluwer – România, București, 2011. [The current legal regulation of banking activity in “Legal aspects in a changing banking system”, Wolters Kluwer – Romania, Bucharest, 2011;] 15. ECB Opinion on the establishment of a consumer protection regime within the scope of the Bank of Greece’s prudential supervisory role and other related provisions (CON/2006/38), Greece, 25 July 2006, https://www.ecb.europa.eu/ecb/legal/date/2006/html/act_8074_amend.en.html 16. ECB Opinion on the establishment of a consumer protection regime within the scope of the Banco de Portugal's prudential supervisory role (CON/2007/29), Portugal, 5 October 2007, https://www.ecb.europa.eu/ecb/legal/opinions/html/act_9727_amend.en.html 17. ECB Opinion on the independence of Banka Slovenije and of the members of its decision-making bodies (CON/2015/8), Slovenia, 2 March 2015, https://www.ecb.europa.eu/ecb/legal/date/2015/html/index.en.html 18. ECB Opinion on the credit mediator legal framework (CON/2015/12), Portugal, 24 March 2015, https://www.ecb.europa.eu/ecb/legal/date/2015/html/index.en.html 19. ECB Opinion on the legal framework for the deposit guarantee scheme and resolution in the financial markets (CON/2015/17), Poland, 28 May 2015, https://www.ecb.europa.eu/ecb/legal/date/2015/html/index.en.html 20.ECB Opinion on certain Česká národní banka tasks in the area of consumer protection (CON/2007/8), Czech Republic, 21 March 2007, https://www.ecb.europa.eu/ecb/legal/date/2007/html/act_9537_amend.en.html 21. ECB Opinion on the role of Národná banka Slovenska in the resolution in the financial market (CON/2015/3), Slovakia, 21 January 2015, https://www.ecb.europa.eu/ecb/legal/date/2015/html/index.en.html 22. ECB Opinion on the legal framework for Narodowy Bank Polski (CON/2015/9), Poland, 10 March 2015, https://www.ecb.europa.eu/ecb/legal/date/2015/html/index.en.html 23. Convergence Report, June 2013, https://www.ecb.europa.eu/pub/pdf/conrep/cr201306en.pdf

88 Central Bank Journal of Law and Finance, No. 1/2015

The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

Mirela Iovu*

Abstract

The objective of Recommendation (EC) 2014/135/EU on a new approach to business failure and insolvency is to ensure that, regardless of their location in the Union, viable enterprises in financial difficulties have access to national insolvency frameworks, which enable the activity restructure at an early stage with a view to preventing their insolvency and, therefore, to maximizing the total value for creditors, employees, owners and economy as a whole. The Recommendation also aims to give honest bankrupt entrepreneurs a second chance across the Union.

Keywords a new approach to business, insolvency, the active role of banks in insolvency prevention

JEL Classification: K220

* Attorney at law, holder of an MBA Diploma in financial management awarded by City University of Seattle, USA, currently a PhD student of the Doctoral School of Law within the Faculty of Law from Bucharest. Vice-president of CEC Bank SA since 2008; Vice-president of the Association of Legal Advisers in the Banking-Financial System since 2008.

Central Bank Journal of Law and Finance, No. 1/2015 89 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

1. GENERAL CONSIDERATIONS ON THE EUROPEAN LEGISLATION IN TERMS OF INSOLVENCY

Regulation (EC) No. 1346/2000 on insolvency proceedings, applied as of 31 May 20021, is the regulation on insolvency at European Union level. The regulation has been adopted to overcome aspects of cross-border insolvency and it is applicable in the case in which the debtor has assets in insolvency or creditors in several Member States. This Regulation contains norms of jurisdiction, recognition, applicable law and provides proper coordination of insolvency proceedings opened in more than one Member State, as well as the avoidance of incentives for the parties to transfer assets or judicial proceedings from one Member State to another, in the attempt to obtain a more favourable legal situation ("the search for the most favourable court" or "forum shopping”). Given its cross-border field of application, the Regulation did not harmonize laws in insolvency matters that serve as the legal basis for solving national causes involving insolvency. Thus, there still remain differences between national laws, therefore economic activities can be lost and creditors recover fewer claims than they should, while creditors of different Member States are not treated in the same way. Intending to reform the European legislation on insolvency and to support Member States in regulating a national legal framework which would correspond to the European principles on this matter, both the European Parliament and the European Commission have already carried out numerous studies and analyses of national laws concerning insolvency.

Thus, according to the results of a study2 commissioned by the European Parliament, the differences between the national laws on insolvency can create obstacles, competitive advantages and/or disadvantages and difficulties for enterprises in the European Union, which have cross-border activities or are owned by natural or legal persons from different Member States. The study found that a harmonization of processes in the insolvency field in the Member States of the European Union would increase the efficiency of the insolvency process and the reorganization of enterprises. In turn, this would allow the satisfaction of creditors’ claims to a greater extent in the case in which a decision is made on the winding-up of the assets or on the improvement of the reorganisation of perspectives, by encouraging a greater number of creditors to support restructuring plans. Together, these measures should increase confidence in the commercial and financial sector, as well as the efficiency in the financial infrastructure of the European Union. Based on the study, the European Parliament has concluded that there are certain areas of legislation related to insolvency in which harmonization is useful and possible. However, any additional analysis of the opportuneness of reforming legislation concerning insolvency shall need to take into account the impact on other important law fields.

90 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu

In January 2011, the European Commission studied business3 dynamics. As it was expected, the study could not prove if the legal system (general law or special law) had any impact on the level of entrepreneurship (rate of setting-up businesses, total entrepreneurial activity and the enterprises’ rate of survival). This means that the efficiency of the insolvency and bankruptcy procedures does not depend on the actual type or aspects on which the legal system focuses, but on specific provisions, as well as on out-of-court settlement of disputes, accelerated procedures for small and medium-sized4 enterprises (hereinafter referred to as SMEs), on a system of early warning and other provisions that significantly affect the efficiency of the system. The countries with the best results in economy (productivity, innovation) have an effective legal framework in terms of insolvency, bankruptcy and systems of early warning for enterprises experiencing financial problems. Based on the results of these studies, in November 2011, the European Parliament adopted its Resolution on insolvency5 proceedings, first resorting to a revision of the Regulation on insolvency. The European Parliament also recommended the harmonisation of specific aspects of the national legislation related to insolvency and of the national law on commercial companies. On the 30th of March 2012, the European Commission started a public consultation regarding the modernization of EU norms on insolvency. The following interested parties were invited to share their experience as far as insolvency is concerned, in particular, cross-border insolvency: small and large enterprises, self-employed persons, specialists in the field of insolvency, legal authorities, public authorities, creditors, academics and the public in general. On December 12, 2012, the Commission adopted a report on the application of Council Regulation (EC) No. 1346/2000 . The report concluded that the Regulation functioned well in general but that it would be desirable to improve the application of some provisions in order to enhance the effective administration of cross-border insolvency proceedings. Since that Regulation has been amended several times and further amendments are to be made, it should be recast for clarity purposes. At the same time, in July 2013, the Commission started a public consultation on the European approach to business failure and insolvency in order to collect points of view regarding key issues, such as: granting a "second chance", the period required for discharge of debts, the conditions to initiate proceedings, the norms on restructuring plans and other necessary measures for SMEs. An additional reflection regarding the issue of the "second chance" refers to an entrepreneur who went bankrupt and now resumes work. Member States have submitted plans of reforming the national legislation on insolvency in order to support

Central Bank Journal of Law and Finance, No. 1/2015 91 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency entrepreneurs6 who want to get a second chance. Most national laws do not seem to help entrepreneurs who want to start a business after they failed in business. For these reasons, few enterprises restart their activity, in spite of the fact that entrepreneurs who have failed are strongly prone to resume activity. Based on the analysis of the above findings, the Commission has identified a series of areas in which differences between national laws on insolvency may create legal uncertainty and an unfavourable business environment. This creates a less favourable climate for cross-border investment. Thus, the second principle of the Commission’s communication "Small Business Act for Europe"7 aimed to promote a second chance for honest entrepreneurs8. "Honest bankrupt" refers to the case in which the failure of the enterprise was not caused due to an obvious fault of the owner or manager; he has been honest and fair, unlike the situations in which bankruptcy was caused by a fraud or an irresponsible behaviour. The communication appeals to the carrying out of exchanges of best practices between Member States. The European Union pays special attention to the situation faced by small and medium- sized enterprises and to granting them a second chance. The Commission thinks that small and medium-sized enterprises (SMEs) should benefit from support for economic difficulties: to prevent insolvency; for the timespan after bankruptcy and for the granting of a second chance; extrajudicial proceedings of solving disputes; judicial proceedings. Restructuring can be extremely costly for SMEs, so the only viable option is often bankruptcy. It was therefore concluded that Member States must find solutions to reduce restructuring costs for SMEs and to set financing costs at bearable levels and in accordance with the real possibilities of these enterprises, so that they can recover from the financial point of view and resume payments to creditors and suppliers under normal conditions and in a reasonable timespan. Extrajudicial proceedings for the restructuring of a business should be open to all types of debtors, regardless of their real possibilities of paying debts at a certain moment. Even if the average duration of extrajudicial proceedings of solving disputes is relatively short, the rate of success of these settlements is above 50 percent in most of the Member States of the European Union. Although the extrajudicial proceedings of solving disputes and the proceedings preceding insolvency are recently introduced mechanisms, they are more and more frequently used by SMEs in the European Union. As in a domino effect, SMEs can also be affected by economic difficulties of other SMEs, which, in their turn, are creditors. It can also be concluded based on the study 9 that some representatives of the business environment, including small and medium-sized enterprises (SME) consider that, in their capacity as creditors, they lose an unjustified

92 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu proportion of their receivables within the framework of the insolvency proceedings because of long proceedings and national norms concerning prioritisation in recovering receivables (preference order). Long and expensive insolvency proceedings limit the granting of a second chance to a great extent. In addition, as a general rule, entrepreneurs who went bankrupt honestly shall be subject to the same limitations as those who went bankrupt fraudulently. This does not mean that only honest entrepreneurs who failed in business are dealing with the social stigma associated with bankruptcy. There are also legal and administrative obstacles when resuming economic activity. Difficulties related to obtaining financing for a new business are considered to be the main problem for entrepreneurs who want to resume economic activity, but it should also be borne in mind that the ones who are trying to recommence an economic activity learn from their own mistakes and they usually record growth faster than newly established enterprises.

To this end, the European Commission recommends10 that Member States should adopt measures in order to make a clear distinction between honest and fraudulent bankruptcy. Legislation on insolvency could distinguish between debtors whose actions or behaviour in business was straightforward, but led to debt, and those who acted unfairly. Legislation could, for example, contain a provision according to which a debtor’s deliberate or irresponsible non-compliance with legal obligations is subject to civil or criminal penalties, as appropriate. Any support programmes for establishing a new business should only be available for honest entrepreneurs even if they went bankrupt, without any distinction between them and the enterprises without financial difficulties. As the European Commission sees it, the following measures are the most important in order to get a second chance: . Winding-up proceedings for honest entrepreneurs, which should be different from those for dishonest entrepreneurs; . Accelerated preparation and implementation of winding-up proceedings for businesses that went bankrupt honestly. In addition, the discharge of obligations is, in turn, essential for granting a second chance: a period of three years to discharge obligations and to pay their debts should be a reasonable maximal time limit for honest entrepreneurs and the procedure should be as simple as possible. It is essential for the entrepreneurial spirit not to become a "conviction for life" when entrepreneurs deal with failure.11 Member States agreed on the necessity to harmonize "the necessary time to discharge obligations" to less than three years. The shortening and alignment of "the necessary period to discharge obligations" would constitute a major step forward in creating a more

Central Bank Journal of Law and Finance, No. 1/2015 93 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency accessible and more innovative business environment, which would enable European enterprises to operate on an equal footing. Moreover, this alignment could represent a first step towards a greater harmonisation of national laws on bankruptcy. Recently, the European Parliament and the Council have adopted the Regulation (EU) 2015/848 of May 20, 2015 on insolvency proceedings12 that will abrogate the Regulation (EC) No. 1346/2000 effective from June 26, 2015. The provisions of this new Regulation shall apply only to insolvency proceedings opened after June 26, 2017. Acts committed by a debtor before that date shall continue to be governed by the law which was applicable at the time they were committed. Notwithstanding some exceptions provided by Article 9213 of the new Regulation, Regulation (EC) No. 1346/2000 shall continue to apply to insolvency proceedings which fall within the scope of that Regulation and which are opened before June 26, 2017. The scope of this Regulation should extend to proceedings which promote the rescue of economically viable but distressed businesses and which give a second chance to entrepreneurs. It should, in particular, extend to proceedings which provide for the restructuring of a debtor at a stage when there is only a likelihood of insolvency, and to proceedings which leave the debtor fully or partially in control of its assets and affairs. It should also extend to proceedings providing for a debt discharge or a debt adjustment in relation to consumers and self-employed persons, for example by reducing the amount to be paid by the debtor or by extending the payment period granted to the debtor. Since such proceedings do not necessarily entail the appointment of an insolvency practitioner, they should be covered by this Regulation if they take place under the control or supervision of a court. In this context, the term ‘control’ should include situations in which the court only intervenes if requested by a creditor or other interested parties.

2. GENERAL CONSIDERATIONS ON THE EUROPEAN COMMISSION RECOMMENDATION (EC) NO. 135/2014 REGARDING A NEW APPROACH TO BUSINESS FAILURE AND INSOLVENCY

Following the consultation from July 2013, the European Commission adopted, on March 12, 2014, Recommendation No. 135 "on a new approach to business failure and insolvency"14, which aims to encourage Member States to set up a framework, a number of common principles for insolvency and pre-insolvency proceedings, to enable the effective restructuring of viable enterprises facing financial difficulties and to provide honest entrepreneurs with a second chance. Thus, it is desirable to promote entrepreneurship15, investment and employment in order to contribute to the reduction of obstacles for the proper functioning of the internal market.

94 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu

Recommendation No. 135 attempts to restructure debtor's activity in order to avoid insolvency and bankruptcy, and when insolvency cannot be avoided, the debtor should be able to adopt a plan of judicial reorganization that would lead to the salvation of the activity. Thus, at the European Union level, approximately half of the firms have a life expectancy of less than five years and about 200,000 companies become insolvent each year, namely 600 enterprises every day, and almost 1.7 million people are likely to lose their jobs each year as a result of this situation. A quarter of these bankruptcies have a cross-border basis and their number is increasing - has doubled from the beginning of the crisis and this trend was maintained in 201416. As a result of the aforementioned findings, it becomes clear that viable measures must be taken at the European Union level in order to prevent or stop the insolvency process and to provide viable businesses with the opportunity to benefit from restructuring and to stay on the market. This is the declared purpose of European Commission Recommendation No. 135 of 2014.

In the light of the experience of some Member States of the European Union17, the companies in difficulty have the opportunity to be restructured earlier and their salvation chances are bigger. However, early restructuring (before the official launch of the insolvency proceedings) is not possible in many countries (for example, in Bulgaria, Hungary, the Czech Republic, Lithuania, Slovakia, Denmark), and if it is possible, proceedings may be ineffective or costly, reducing the incentives offered to enterprises to keep them afloat. Finally, in some countries many years may pass until honest entrepreneurs who went bankrupt can benefit from the discharge of old debts and can test a new business idea (Austria, Belgium, Estonia, Greece, Italy, Latvia, Lithuania, Luxembourg, Malta, Croatia, Poland, Portugal and Romania). By shortening the length of the debts discharge when an honest entrepreneur goes bankrupt we would make sure that bankruptcy does not become a "conviction for life". Recommendation No. 135 of 2014 of the European Commission is trying to homogenize the rules in different Member States of the European Union, in the insolvency field, thereby giving chances for recovery and cross-border groups of companies. A more coherent approach at the European Union level would not only improve the level of amounts recovered by creditors and the cross-border investments flow, but it would also have a positive impact on the entrepreneurial spirit, on employment and innovation18. According to the Recommendation, the Member States of the European Union must give effect to the principles set out in the text of the recommendation, through the establishment of the appropriate measures, at legislative level, until March 14, 2015. European Commission’s evaluation of the Recommendation implementation is going to

Central Bank Journal of Law and Finance, No. 1/2015 95 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency be carried out within a period of maximum 18 months, namely until September 14, 2015. Thus, after 18 months from the adoption of the recommendation, the Commission will carry out the assessment of the state of affairs, based on the Member States’ annual reports, in order to determine whether additional measures are necessary to consolidate and strengthen the approach reflected in the recommendation.

3. THE NATIONAL IMPLEMENTATION OF EUROPEAN COMMISSION (EC) RECOMMENDATION NO. 135/2014 ON A NEW APPROACH TO BUSINESS FAILURE AND INSOLVENCY

The implementation of the Commission’s Recommendation at national level should lead to the establishment of a coherent framework for national legal norms on insolvency prevention. According to Recommendation No. 135/2014 of the European Commission, as a Member State of the European Union in which the average of businesses in bankruptcy is relatively high in comparison with the European average, Romania should be constantly concerned with the implementation of the principles contained in the recommendation. Thus, the restructuring procedure or the judicial reorganization of an undertaking in difficulty should not be long-termed and expensive; it should be flexible, so that more steps can be taken out of the court; it should also be based on a realistic and feasible restructuring plan. Like all the other Member States, Romania too must make sure that courts are in the position to confirm the reorganisation plans rapidly and, basically, by means of a uniform procedure that sets clear and specific provisions concerning the content of the restructuring plans as well, not only from the procedural point of view or from the standpoint of the approval mode and plan confirmation. In practice, the reorganisation plans are usually imposed by debtors and creditors; the latter have very little time to analyse them or they are not allowed to propose changes before submitting them to formal approval, which leads to the rejection of plans, even in the case of plans that would have had chances of successful implementation. Act no. 381/2009 on the scheme of preventive composition (conciliation procedure) and the ad hoc mandate, as published in Monitorul Oficial, Part I, No. 870 of December 14, 2009, as subsequently amended, has brought to Romania the pre-insolvency proceedings. The law subsequently amended has been integrated into Law No. 85/2014 on insolvency prevention and insolvency proceedings. Such pre-insolvency arrangements consist in the ad hoc mandate and the scheme of preventive composition with the creditors (or preventive concordat). The ad hoc mandate

96 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu is a confidential proceeding, initiated by the debtor who finds himself in a difficult financial situation. In this procedure, the court appointed ad hoc administrator negotiates with the creditors to reach a settlement between at least one creditor and the debtor, in order to help the debtor get out of its difficult situation. The preventive scheme of composition [conciliation procedure] is a confidential procedure, initiated by the ailing debtor. A court appointed conciliator negotiates with creditors, to reach an arrangement between all or most of them and the debtor, in order to help the debtor out of its financial predicament. This procedure has broader effects than the ad hoc mandate. If ratified, the scheme of composition is erga omnes, not just binding the creditors who signed the agreement. The preventive composition, as a contractual arrangement, is an agreement signed by the debtor on the one hand and the creditors who own at least two thirds of the value of the approved and unchallenged claims on the other hand. The debtor proposes a plan to regain its business and pay back the debts owed to creditors and the creditors agree to support the debtor in its efforts to overcome its financial difficulties. Currently, the ad hoc mandate, as a pre-insolvency proceeding, has no effect. The Romanian internal regulations gave the ad hoc administrator too few powers and actual tools to be able to attract the creditors into negotiations that lead to a significant result, namely an out-of-court settlement between the debtor and its creditors. Therefore, in practice, this pre-insolvency proceeding is rarely chosen19. As to the scheme of composition, this is a procedure regulated by heavy and quite rigid legal dispositions. On the one hand, the hearings are too many and set too far apart, and on the other hand the requirements to ratify the arrangement are too many and too demanding (as the percentage required from the total creditors’ mass) to yield actual results. There are not enough statistical data. The effects of implementing Commission’s Recommendation at national level should lead to the establishment of a coherent framework for national norms in the field of insolvency. The first step in implementing European Commission’s Recommendation into Romanian legislation is represented by the fact that Law No. 85/2014 on insolvency prevention and insolvency proceedings came into force on June 28, 201420. Law No. 85/2014 takes over and implements a part of the provisions from the European Commission’s Recommendation No. 135 of 2014 related to pre - insolvency proceedings, i.e. the ad-hoc term and preventive concordat, and to insolvency proceedings. Applied effectively and pro-actively, the rules concerning the ad-hoc term and the preventive concordat could allow the debtor’s reorganization in order to save the business, prior to the initiation of the insolvency proceedings.

Central Bank Journal of Law and Finance, No. 1/2015 97 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

For instance, in the case of the preventive concordat, the foreclosure suspension principle is implemented, as provided by European Commission Recommendation No. 135 of 2014, both at the request of the debtor before the ratification of the concordat and by right after the ratification of the concordat. The predicted effects of the implementation into the national legislation of the Commission Recommendation (EC) No. 135/2014 are the following: . Activity restructuring of the companies that are in an early stage of financial difficulty, preferably before the opening of the insolvency procedure, in order to avoid the accumulation of substantial costs involved by the procedure, as well as to avoid the liquidation of the business . Activity restructuring without the need to open judicial proceedings officially; . Facilitation of the reorganization plan adoption, taking into account both debtors’ and creditors’ interests, so as to increase the chances of saving viable businesses; . Reduction of the negative impact that bankruptcy would have on the future opportunities of an entrepreneur to lay the foundations of a new business; . Discharging the debtor's debts within no more than three years from the beginning of the reorganisation. The benefits of implementing European Commission Recommendation (EC) No. 135/2014 into the national legislation will become visible along with the effects of applying Law No. 85/2014 on insolvency prevention and insolvency proceedings to a large number of companies saved from bankruptcy through preventive measures or judicial reorganisation successfully implemented.

4. EFFECTS OF IMPLEMENTING EUROPEAN COMMISSION RECOMMENDATION (EC) NO. 135/2014 ON A NEW APPROACH TO BUSINESS FAILURE AND INSOLVENCY IN ROMANIAN BUSINESS ENVIRONMENT

Even though the economic activity in Romania experienced some improvement in 2013 and 2014 compared to the previous years, considered years of crisis, the statistical data from the National Trade Register Office show that, between 2008 and 2013, the number of insolvencies followed an upward trend, reaching to more than 29,500 companies in 2013, 10.7 percent higher than the data recorded in 2012. The number of companies in insolvency decreased by 30 percent in 2014 compared to 2013, namely from 29,587 to 20,696 companies, and the number of insolvent companies in January 2015 was 54.72 percent lower than the one from January 2014, according to the data provided by the

98 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu

National Trade Register Office. Consequently, Romania enjoyed a much lower number of insolvencies in the first quarter of 2015 than in the first quarter of the previous year, due to the introduction of the New Insolvency Act in June 2014. The solid economic activity, supported by stronger household consumption and increased utilisation of EU funds, has also been translated into improvements on the corporate side: insolvencies decreased by 28 percent. Furthermore, the insolvency of 29,500 enterprises in 2013 determined the loss of about 100,000 jobs, more than 60 percent of which were concentrated in the following five sectors: textile products manufacturing, clothing and footwear, construction, chemical substances and products, metallurgical industry. The social impact of insolvent companies in 2013 is similar to the one recorded in 2012, but 41 percent higher than in 2011. According to the same statistics, companies with impact and assets exceeding 1 million euros each have more than 1.87 million employees, i.e. more than one-third of the population employed at national level. Three-quarters of the largest 10 companies that have become insolvent in the last four years are or were controlled by Romanian businessmen, while the rest are held by public authorities or various foreign investors. It seems that successful multinational companies have resisted better during the economic crisis from the past four years, in most cases due to the foreign shareholders’ support. Moreover, the small and medium-sized enterprises are the most affected, which also puts pressure on large and very large companies. Thus, the number of companies, with a turnover higher than 1 million Euros, which became insolvent in the first quarter of 2014, is equal to 188,948, 2.5 times higher than the number recorded in the same period of the previous year. The most significant 10 Romanian companies that became insolvent in the first half of the year 2014 had a level of aggregated turnover of 660 million Euros and an average duration of 16 years, according to a study conducted by the company for risk assessment. It is more alarming that the number of companies in difficulty or imminent insolvency is much higher than the one of sound companies and the number of businesses with impact which are likely to become insolvent is greater than the number of those in the same category which have become insolvent in the last four years. It is important to emphasize that more than 90 percent of the initiated insolvency proceedings lead to bankruptcy because no restructuring plan is approved or, even if such a plan is approved, its provisions are not met, so the number of bankruptcy proceedings is well above the EU average. Taking into account the number of insolvency proceedings initiated in 2013 in the countries from East-Central Europe, Romania stands out with the second highest rate of

Central Bank Journal of Law and Finance, No. 1/2015 99 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency insolvencies related to the number of active companies, i.e. 6.44 percent. The only country with a higher percentage is Serbia, where the percentage of 7.61 percent is rather due to a base effect (a low percentage of active companies). Moreover, the number of insolvencies initiated in 2013 is more than twice the level registered in Hungary, although the number of active firms (with a turnover greater than zero) is 30 percent lower in Romania. The comparison becomes even more eloquent when analysing the figures recorded in Romania compared to Poland, which records 300 times fewer insolvent companies (only 818 newly opened insolvency proceedings in 2013) and a 4 times higher number of active companies in comparison with Romania. The insolvencies phenomenon has developed locally, especially during the post-crisis period of the last 5 years, with Romania generating approximately 40 percent of the total insolvencies opened during the year 2013 in East-Central Europe, at the same time recording only 6 percent of the total active companies in the region. In Romania, this phenomenon is amplified by a very permissible legislative framework for the debit party, which encourages the insolvency proceeding as an effective measure of protection against creditors.21 If before the onset of the crisis only one firm out of 30 became insolvent, in 2013 things completely changed, in the sense that one company out of seven was insolvent. All these companies have credits from banks in Romania worth 3.7 billion Euros and there is a great uncertainty regarding their recovery. All statistics show that, best case scenario i.e. that of guaranteed creditors, they are likely to recover approximately 30 percent of the claims from the statement of affairs, in approximately 3.3 years22 from the initiation date, namely the average duration of an insolvency procedure. Banks have concentrated their efforts on restructuring loans granted to companies in difficulty by: debt rescheduling arrangements, granting a grace period, extending maturities, reducing costs temporarily, which can bring in a relative balance in the debtor companies’ balance sheets, useful for the restructuring/reorganization of the company and for insolvency prevention. We believe banks can play a more active role than that. For example, the restructuring plans of companies in reorganisation could include, based on partnerships with banks, financing strategies for the companies’ suppliers, distributors and customers, with a synergetic effect on all those involved in the vertical gear. In addition, because of the involvement in the "pre-packed" procedure (or "controlled” insolvency arrangement) which offers the opportunity of an agreement between the debtor and the big creditors (banks, the big suppliers, budget creditors) of a company in difficulty, before the official declaration of insolvency, such an agreement should come into force immediately after

100 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu the court order regarding the opening of proceedings and the appointment of the judicial manager.

Therefore, in Romania the "insolvency" phenomenon is considered a systemic risk23 because there have been identified more structural causes fuelling it: . The increasing importance of commercial credit, on the background of a higher share of claims in current assets, in accordance with corporate balance sheets; . Time limits extension for collections, given the more aggressive growth in claims, compared to the dynamics of the turnover (business relationships with partners presenting a higher risk) - firms that became insolvent in 2013 had shown a decrease of 11 percent in turnover in 2012 compared to 2011, while the amount of claims had increased by 2 percent; . The level of arrears still high and the negative impact they have on businesses working with the state; . The companies’ more precarious situation from a financial point of view and prospects of self-financing, given their own eroded reserves and as a result of the financing restrictions that have persisted in the last 3 - 4 years. It appears that, in the case of companies that had become insolvent before 2013, it was not the negative financial performance that put an end to their activity (it surely mattered, but it did not play a crucial role in determining insolvency), but rather the faulty decisions of credit risk management, the net cash flow management and the appropriate funding of the operation cycle, the lack of capitalisation and the continuous process of disinvestment in certain companies. All these causes generate critical consequences for the economy because a large part of the big Romanian companies, on which the re-launching of the economy depends, have come on the brink of bankruptcy. Thus, of the total of over 120,000 companies that have become insolvent during the last 4 years, 2,200 are considered companies with a significant impact on the economy. The only conclusion we can draw from such gloomy statistics is the following: if the banking system does not take on a more active role in the financial restructuring of companies' debts and without a continuous improvement of the national legislation in the field of insolvency prevention with a greater emphasis on prevention procedures, the number of insolvency cases registered on the territory of Romania will continue to increase, affecting both large companies and SMEs.

Central Bank Journal of Law and Finance, No. 1/2015 101 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

5. THE EFFECTS OF IMPLEMENTING EUROPEAN COMMISSION RECOMMENDATION (EC) NO. 135/2014 ON A NEW APPROACH TO BUSINESS FAILURE AND TO INSOLVENCY IN BANKING ENVIRONMENT

The reform of national norms on insolvency prevention would be for the benefit of all interested parties (professionals (enterprises), creditors, employees, shareholders and associates) because it would allow viable companies to continue their activities and to protect jobs. At the same time, it would create a more favourable environment for creditor institutions, which would thus be able to recover a larger part of investments than they would in the case in which the debtor companies were to go bankrupt.24 The effects of insolvency on banking environment are obvious, consisting mainly in the rate increase for non-performing credits up to 22 percent of the banking system level, on 31 March 2014. Before the entry into force of Law No. 85/2014 and the implementation of European Commission Recommendation No. 135 of 14 March 2014, banks’ exposure on insolvent firms stood at the level of 5.5 billion euros of which approximately 1 percent was to be recovered25. The need for adopting a new law on insolvency was obvious. The improvement of the old legislation was to bring advantages to several areas, but in particular to the banking system. The solutions to these problems come from intervening on time, from finding some financing sources, from the opportunities for restructuring and from rethinking crisis management. Law No. 85/2014 transposes part of the principles set out in European Commission Recommendation No. 135 of 2014 into national law, especially through the creation of pre-insolvency leverage to recover companies. To this end, we should mention one of the principles laid down in Article 4 of Law No. 85/2014: “Providing borrowers with an actual and effective chance to recover their business, either through insolvency prevention proceedings, or by means of the judicial reorganization procedure." This principle is implemented by the provisions related to the ad-hoc term and to the preventive concordat. We think the implementation of Recommendation No. 135 of 2014 will have a positive impact on the banking field, both on crediting and on the possibilities of recovering claims, leading to more efficient pre - insolvency and insolvency proceedings. Most banks in Romania have internal norms for the restructuring of flexible and permissive credits, in order to restore confidence and supply their clients with the necessary "fuel" to continue the economic activity. It is normal for such behaviour of banks to exist and it is unfair for a bank not to save a customer when in a difficult situation due to the narrowing or

102 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu disappearance of outlets, thus acting according to the principle "One gives you an umbrella when it's sunny and takes it away when it’s raining". Moreover, we anticipate that the duration of the insolvency proceeding will decrease, making it more efficient, there will be fewer disputes within the procedure and debtors will be offered a real opportunity for reorganization by conducting insolvency proceedings as set out by the framework of Law No. 85/2014 on insolvency prevention and insolvency proceedings.

6. FUTURE LAW PROPOSALS FOR THE IMPLEMENTATION OF THE EUROPEAN COMMISSION RECOMMENDATION (EC) NO. 135/2014 ON A NEW APPROACH TO BUSINESS FAILURE AND INSOLVENCY

By adopting Law No. 85/2014 on insolvency prevention and insolvency proceedings, important steps were made towards the implementation of European Commission Recommendation (EC) No. 135/2014. However, the Recommendation also offers other possibilities of improvement and consolidation of the legislation in force in order to provide the companies in financial difficulty with more chances for recovery. We believe that an improvement in the current legislative framework could be attained through the differentiation of the reorganization plan duration according to the debtor’s particularities (SMEs / companies with high turnovers / groups of entities), as well as through a more efficient "honest" insolvency with an emphasis on judicial reorganization at the expense of the "fraudulent" one. Another proposal of future law could refer to the faster discharge of debt, which may grant quick chances of economic recovery to the honest debtor. As an additional proposal of future law, the increase in the threshold debts’ amount from 40,000 lei26 (as it is currently) to 100.000 lei could be taken into consideration, because it would offer the possibility of introducing the request concerning insolvency opening. In our opinion, by significantly increasing the threshold amount, creditors and debtors would be motivated to turn to pre-insolvency proceedings before reaching a debt of such an amount. Thus, the number of companies in insolvency will diminish and the possibilities of "a second chance" increase.

Central Bank Journal of Law and Finance, No. 1/2015 103 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

NOTES

1 Regulation (EC) No. 1346/2000 of the Council on insolvency proceedings was published in the OJ L 160. 30 June 2000. 2 "Harmonisation of legislation in the field of insolvency at EU level, European Parliament 2010, IN 419,633. This was followed by the study "harmonisation of legislation in the field of insolvency at EU level in relation to the opening of proceedings, the declaration and verification of claims, and the reorganisation plans", IN 2011, TO 432,766. 3 "Business dynamics: start-ups, business transfers and bankruptcy" (Dynamics of firms: newly set- up businesses, transfer of undertakings and bankruptcy), European Commission, DG Enterprise and Industry, January 2011. This report contains a study regarding the economic impact of administrative and judicial proceedings related to bankruptcy and to the opportunities for a second chance after bankruptcy in 33 European countries (27 EU Member States, plus Iceland, Croatia, Norway, Turkey, Serbia and Montenegro). 4 SME is an acronym for small and medium-sized enterprises, as defined in EU legislation, namely Recommendation 2003/361/EC of the Commission, published in the EU Journal L124, p. 36 of 20 May 2003; the criteria by which SMEs are distinguished from other types of undertakings are: (i) Number of employees and (ii) turnover or total balance: Type of enterprise: Number of employees Turnover or Total balance Average < 250 ≤ €50 million ≤ €43 million Small < 50 ≤ €10 million ≤ €10 million Micro < 10 ≤ €2 million ≤ €2 million

5 Resolution EP of 15 November 2011 containing recommendations for the Commission on insolvency proceedings in the context of the EU law on commercial companies. 6 Romania adopted Law No. 85/2014 on insolvency prevention and insolvency proceedings, which was published in Monitorul Oficial al României 466/25 June 2014 and entered into force on 28 June 2014. 7 The Commission’s Communication to the European Parliament, the Council, the Economic and Social Committee and the Committee of Regions of June 25, 2008, entitled 'think on a small scale first": priority for SMEs - a "Small Business Act" for Europe" COM (2008) 394 final - Not published in the Official Journal). 8 Principle II: "Provide the possibility for honest contractors who went bankrupt to benefit from a second chance in a fast way". 9 Business dynamics: start-ups, business transfers and bankruptcy" (Enterprises dynamics: newly founded businesses, transfer of undertakings and bankruptcy), European Commission, DG Enterprise and Industry, January 2011. 10 European Commission Recommendation No. 135/2014 concerning a new approach to business failure and to insolvency of March 12, 2014, published in the Official Journal of the European Union No. 74 of March 14, 2014.

104 Central Bank Journal of Law and Finance, No. 1/2015

Mirela Iovu

11 This was one of the recommendations in the panel of experts’ report on a second chance. 12 the Regulation (EU) 2015/848 of May 20, 2015 on insolvency proceedings has been published in the Official Journal of the European Union No. L 141/19 of June 5, 2015. 13 Article 92-Entry into force. This Regulation shall enter into force on the twentieth day following that of its publication in the Official Journal of the European Union. It shall apply from 26 June 2017, with the exception of: (a) Article 86, which shall apply from June 26, 2016; (Information on national and Union insolvency law 1. Member States shall provide, within the framework of the European Judicial Network in civil and commercial matters established by Council Decision 2001/470/EC (17), and with a view to making the information available to the public, a short description of their national legislation and procedures related to insolvency, in particular to the matters listed in Article 7 (2). 2. Member States shall update the information referred to in paragraph 1 regularly. 3. The Commission shall make information concerning this Regulation available to the public). (b) Article 24(1), which shall apply from 26 June 2018; and (Establishment of insolvency registers 1. Member States shall establish and maintain in their territory one or several registers in which information concerning insolvency proceedings is published (‘insolvency registers’). That information shall be published as soon as possible after the opening of such proceedings). (c) Article 25, which shall apply from June 26, 2019. (Interconnection of insolvency registers 1. The Commission shall establish a decentralised system for the interconnection of insolvency registers by means of implementing acts. That system shall be composed of the insolvency registers and the European e-Justice Portal, which shall serve as a central public electronic access point to information in the system. The system shall provide a search service in all the official languages of the institutions of the Union in order to make available the mandatory information and any other documents or information included in the insolvency registers which the Member States choose to make available through the European e-Justice Portal. 2. By means of implementing acts in accordance with the procedure referred to in Article 87, the Commission shall adopt the following by June 26, 2019: (a) the technical specification defining the methods of communication and information exchange by electronic means on the basis of the established interface specification for the system of interconnection of insolvency registers; (b) the technical measures ensuring the minimum information technology security standards for communication and distribution of information within the system of interconnection of insolvency registers; (c) minimum criteria for the search service provided by the European e-Justice Portal based on the information set out in Article 24;

Central Bank Journal of Law and Finance, No. 1/2015 105 The Effects of Implementing European Commission Recommendation (EC) No. 135/2014 on a New Approach to Business Failure and Insolvency

(d) minimum criteria for the presentation of the results of such searches based on the information set out in Article 24; (e) the means and the technical conditions of availability of services provided by the system of interconnection; and (f) a glossary containing a basic explanation of the national insolvency proceedings listed in Annex A).

14 European Commission Recommendation No. 135/2014 concerning a new approach to business failure and insolvency of March 12, 2014, published in the Official Journal of the European Union No. 74 of March 14, 2014 15 On January 9, 2013, the Commission adopted the action plan "Entrepreneurship 2020 ", through which Member States are invited, among other things, to reduce, where possible, the period required to discharge obligations and to recover debts, in the case of honest entrepreneurs who went bankrupt, to a maximum of three years, until 2013, and to provide support services for the early restructuring of firms, as well as counselling to avoid bankruptcy and support for restructuring and re-launching small and medium-sized enterprises. 16 http://ec.europa.eu/romania/news/12032014_comisia_europeană_proceduri_de_insolvenţă_ro.htm 17 “Business dynamics: start-ups, business transfers and bankruptcy" (Enterprises dynamics: newly set-up businesses, transfer of undertakings and bankruptcy), European Commission, DG Enterprises and Industry, January 2011 18 The press release issued by the European Commission - Brussels March 12, 2014 19 “Study on a new approach to business failure and insolvency – Comparative legal analysis of the Member States’ relevant provisions and practices”, INSOL Europe Report for EC 2014 Annex 1 - European Commission, www. ec.europa.eu/justice/civil/files/insol_europe_report_2014_annexes_en.pdf 20 Law No. 85/2014 on insolvency prevention and insolvency proceedings, published in Monitorul Oficial al României No. 466 of 25 June 2014 21 Insolvency Report 2013 for EEC, Coface Romania, http://www.coface.ro/Ştiri- Publicaţii/Ştiri/Raportul-de-Insolvenţă-2013-pentru-CEE-Numeroase-insolvenţe-din-cauza-unui- mediu-economic-neprielnic 22 World Bank; www.doingbusiness.org 23 Report on financial stability - 2014, National Bank of Romania, http://www.bnro.ro/PublicationDocuments. 24 The press release issued by the European Commission – Brussels, March 12, 2014 25 Financial Newspaper Supplement, March 27, 2014 26 Article 5, Point 72 - threshold amount, of Law No. 85/2014 on insolvency prevention and insolvency proceedings.

106 Central Bank Journal of Law and Finance, No. 1/2015 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

Irina Mihai* Alina Tarța**

Abstract

The corporate insolvency framework is an important part of the resolution mechanism for financial distress and its design affects the resources allocation in the economy and the development of the credit market. In this paper we analyse the Romanian framework of the corporate insolvency in the light of the existing literature and assess its implications for the economy and the financial sector. We dwell upon the efficiency of the existing insolvency framework and on the important take-away lessons for policy makers.

Keywords corporate insolvency, bankruptcy, efficiency, bank credit, Romania

JEL Classification: G21, G33, G38

* National Bank of Romania, Financial Stability Department, [email protected]. ** PhD, National Bank of Romania, Cluj Regional Branch, [email protected]. The opinions expressed in this paper are those of the authors and do not necessarily represent the views of the National Bank of Romania, nor do they engage it in any way. The authors would like to thank Elena Banu and Ruxandra Popescu for their help with processing data and for their useful comments.

Central Bank Journal of Law and Finance, No. 1/2015 107 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

1. INTRODUCTION

Insolvency is an important process for the economy, necessary for economic growth. It affects the resource allocation in the economy by providing a framework for solving financial distress and for transferring assets away from unviable businesses. The insolvency procedure creates additional value to the direct liquidation of the firm as it requires consensus among creditors which should offer protection to all the creditors of the firm, increase the transparency of the process and allow for the viable firms to continue their activity. Another mechanism through which corporate assets are reallocated in the economy, ideally for a more efficient use, is by mergers and acquisitions (M&A), as firms’ competitors decide to buy the firm in financial distress to acquire its assets and/or its clients (Dahiya and Klapper, 2007). This second mechanism is less used in the Eastern European countries, and even less in Romania, as the market for M&A is less developed. In the last 8 years (2007-2014), in these countries, M&A represented only 15 percent, in average, from all transactions in Europe, while in Romania were registered as much as 4 1 percent from all mergers and acquisitions in the region .

The insolvency framework plays a higher role during economic downturns and early recovery phase of the economic cycle. Studies that followed the 1997-1999 financial crises found strong evidence for the link between the quality of the insolvency framework and the speed of economic recovery (Hussain et al. 1999, Claessens et al. 2005). The weaknesses in the corporate insolvency system are monitored by the World Bank to assess the regulatory environment from the perspective of easiness to do business in a country. The index of insolvency measures the duration, the cost, the outcome and the strength of the insolvency procedures. The financial crisis tested the corporate insolvency mechanism, as the number of companies that filed for insolvency rose immediately after the crisis in almost all countries monitored by the World Bank. The recent data indicates that while the number of new firms under insolvency proceedings is decreasing, the level is still high compared to the pre-crisis levels. The efficiency of the insolvency process is assessed ex-ante, as the incentives for debtors and shareholders to delay filing for bankruptcy and to engage in riskier or even speculative activities (measuring the ability of the legal framework to limit moral hazard), and, ex-post, as the debt recovery rate or as total costs incurred by the firms under the insolvency procedure.

1 According to the Thomson Reuters database on mergers and acquisitions.

108 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

There is no consensus on what should the optimal design of corporate insolvency framework be. The way the insolvency procedures are designed can either favor debtors or creditors. The debtor friendly bankruptcy frameworks encourage entrepreneurship but might give support to unprofitable business and promote riskier business decisions as debtors can exit an unprofitable business with minimum cost. On the other hand, the creditor friendly frameworks are found to help the development of credit market, while the limitations of these frameworks are the excessive usage of the liquidation mechanism and of assets fire-sales. This type of system might also incentivize debtors to delay filing for bankruptcy and to engage in risky projects to obtain higher returns needed to compensate the accumulated losses. The literature finds a strong link between countries’ legal origin and the level of creditor rights, Djankov et al. (2007). Civil law countries (especially French civil law countries) rely on government regulation and have a much higher reliance on public credit registries (Djankov et al., 2007), while common law countries use private contracting (Djankov et al., 2003 and Mulligan and Schleifer, 2005) and have much higher creditor rights. Jappelli et al. (2002) suggest that higher creditor rights and better (ex-ante and interim) information are substitutes, as civil law countries develop information institutions, while the others extend the legal systems to give more power to creditors. Djankov et al. (2007) argue that both better creditor rights and the presence of credit registries are associated with credit market development and higher private credit relative to GDP, and finds evidence on the improvements in credit markets after reforms that either increase creditor rights or introduce credit registries. On the other hand, Biais et al. (2004) suggest that soft laws, while reducing the liquidation frequency, they might increase ex-ante credit rationing. Garcia-Posada (2013) evaluates the Spanish insolvency framework from the perspective of incentives in the credit market and finds it as inefficient. This is mainly due to the existence of an alternative system to insolvency that banks can use in order to recover the value of their claims and which is more creditor-friendly. This second mechanism, which is the mortgage system, is extensively used by banks and stimulates firms to overinvest in fixed capital in order to reduce their funding cost, thus contributing to a misallocation of assets in the economy. As a consequence, the number of formal bankruptcy applications is very low, even during deep economic downturns. Davydenko et al. (2008) find evidence in France, where banks responded to a less creditor-friendly framework by requiring higher collateral on firms’ credit. Tilting the insolvency law more in debtor’s or creditor’s favor may not be the relevant question, as the predictability of how the rules are applied and the transparency of the process appear to matter more (Hagan, 2000).

Central Bank Journal of Law and Finance, No. 1/2015 109 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

A general accepted idea is that an important factor of the strength of the insolvency framework is the efficiency of the judicial system (Padilla et al. 2000). Studies have found that in economies with weak judicial systems, debtors are more likely to exhibit risky behavior, resulting in higher number of bankruptcy applications and increased level of financial distress (Claessens et al. 2005). In this paper we analyse the Romanian framework of the corporate insolvency with a focus on its impact on both the payment discipline in the economy and the credit market. We also dwell upon the efficiency of the existing insolvency framework and on the important take-away lessons for policy makers. In this view, the rest of paper is organized as follows. Section 2 presents the current amplitude and the dynamics of the insolvency phenomenon in Romania. Section 3 discusses the efficiency of the insolvency process as a resolution mechanism for financial distress firms and section 4 concludes.

2. THE CORPORATE INSOLVENCY IN ROMANIA

The corporate insolvency is an important process in Romania. Between 2010 and 2014, around 30 percent of the exiting firms were liquidated through the insolvency procedure (Table 1, Annex). Compared to other EU states, Romania has one of the highest utilization rates of the insolvency procedure. According to Creditreform Economic Research Unit data, Romania was ranked, in 2012, the fourth highest among EU member states, measured in both absolute and relative terms. Even comparing to countries with the same legal origin, Romania has the highest degree of reliance on this procedure (Graph 1, Annex). The number of companies filing for insolvency grew significantly after the recent financial crisis. The number of new insolvency cases increased by 65 percent in 2013, compared with 2008. In 2014, the number decreased significantly, signalling a possible trend shift in the amplitude of the insolvency phenomenon (Table 1, Annex). Despite this reduction 2 in the intensity of the insolvency process (the density of new firms under insolvency proceedings declined from 2.2 firms to 1.5), the insolvency phenomenon continues to add to a steady growing gap between newly created firms (with a density of 4.2 firms in 2014) and exiting firms (with a density of 5.6 firms in 2014). An upward trend in filing for insolvency was observed in almost all countries, according to the World Bank. The dynamics among developed European economies was more pronounced in countries like Spain, Portugal and Ireland (the number of corporate

2 The density of firms in insolvency is calculated as number of firms under insolvency procedure per 1000 people ages 15-64 years.

110 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța insolvencies grew by more than 100 percent, Graph 3, Annex). The NM-113 countries also recorded significant growth rates, with higher rates between 2010 and 2012 observed in Croatia, Hungary and Bulgaria (Creditreform Economic Research Unit, 2012). The financial crisis tested the corporate insolvency mechanism, urging for reforms. According to the World Bank, there were around 50 reforms of the corporate insolvency legal framework during 2008-2014 in EU countries. Romania was no exception. In 2011, a procedure for out-of-court settlements was introduced. In 2012, the law was amended to reduce the duration of insolvency proceedings. In 2014, the insolvency framework was substantially changed. We will discuss in the last part of this section some of the newly introduced features. The changes made to the insolvency framework will take some time until their impact will be seen, as the largest share of the outstanding companies undergoing insolvency proceedings at the end of 2014 will remain under the incidence of the previous code (Graph 4, Annex). However, this phenomenon is not backed by a high degree of efficiency. According to the World Bank, Romania is in the last 10 EU countries in what concerns the easiness of resolving insolvency (Graph 2, Annex). A firm entering the insolvency procedure in Romania has a high probability of being liquidated. However, starting with 2011, there is an increasing tendency for using insolvency as a debt restructuring framework, as more firms are entering under reorganization procedure. The firms under judicial reorganization represent 5 percent of the new firms that have filed for insolvency or 2 percent of the firms under insolvency procedure in 2014. The number escalated from around 100 in 2009 to over 1000 in 2014. The out-of-court settlements are used to a lesser extent. The number of insolvency procedures cancelled in 2014 was only 364, up from below 300 in 2009 (Table 1, Annex).

2.1. What Firms Undergo Insolvency Proceedings? The companies that have filed for insolvency are mostly companies with Romanian ownership (over 70 percent), established during the expansion phase of the economic cycle (2000-2008, over 50 percent) and are mostly micro companies (86 percent from all companies undergoing insolvency proceedings at the end of the first quarter of 2015). Looking at the economic sectors, the firms under insolvency procedure are mainly from service and trade sectors. The structure of the outstanding firms in insolvency is relatively closely mapped to the structure of the economy with two notable exceptions: (i) the firms established before 2008 are more present in the insolvency pool, and (ii) the firms from industry and agriculture sectors have a larger share of the companies in insolvency (13

3 The NM-11 countries are Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovak Republic and Slovenia.

Central Bank Journal of Law and Finance, No. 1/2015 111 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania percent and, respectively, 8 percent, compared to an average of 7 percent for the whole economy, Graph 5, Annex). This outcome points to another issue of the insolvency mechanism. The companies from industry and agriculture are spending more time under the insolvency procedure compared to companies from trade, services and utilities (25 months in average in the case of companies from industry, compared to 19 months in the case of companies from trade). The companies are usually experiencing financial distress long before entering insolvency, therefore triggering this resolution mechanism is relatively late. More than two thirds of the firms under insolvency procedure have negative capital in the year before declaring insolvency and around one third had negative capital in all the three years before entering under insolvency procedure. If we analyse the companies in the year before entering insolvency, the net profitability of these companies is usually in negative territory while the efficiency of asset utilization is weaker relative to other companies (less by 17 percent, in average, during 2009 - 2013). Moreover, the ability of these firms to collect receivables from their trade partners is feebler compared to the economy. The number of days to collect them is, in their case, around 1.6 higher than for the whole economy (average values for the period 2009-2013, Graph 6, Annex). These companies are confronted, even before entering insolvency, with higher difficulties in accessing finance4, with lesser internal funding sources and with the need to use external funding to finance current activities and even refinance bank credits compared to the whole economy. As external sources, they rely more than the economy on overdraft, working capital credit and on credit from their commercial partners. Another important distinction is that they report, to a higher degree, problems on the supply chain, like payment discipline and/or insolvency of their trade partners (Table 2, Annex). Insolvency is relatively well anticipated by a high indebtedness and a low liquidity ratio. Companies under insolvency procedure have a higher leverage ratio (Graph 7, Annex) and a lower liquidity ratio in the year before declaring insolvency compared to the rest of the economy (Graph 8, Annex).

4 The data is based on a survey of non-financial companies conducted periodically by NBR, Survey on the access to finance of non-financial corporations in Romania and their capacity to withstand adverse financial conditions (FCNEF Survey). The sample for FCNEF Survey is of approximately 10,500 non-financial corporations, it is representative at national and regional levels and it is extracted using statistical procedures, in compliance with the following criteria: i) firm size class (microenterprises, small enterprises, medium-sized enterprises and large companies), ii) economic activity (based on NACE Rev. 2) and ii) development regions. The answers to the survey included in the text are those of companies that were solvent at the time of the survey but enter insolvency after that.

112 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

The direct impact on the economy is relatively limited at the time of declaring insolvency. They contribute only by 4 percent to the gross value added, concentrate 6 percent of the workforce employed in the real sector and hold 9 percent of the total assets of the nonfinancial companies sector. These companies have a modest role for the trade balance, contributing by only 2 percent to the total amount of exports and imports generated by the non-financial companies. As regards the indirect channel, companies undergoing insolvency proceedings have an important role to the payment discipline in the economy. These companies are responsible for a large share of the overdue payments in the economy - one third of the overdue payments to suppliers and around one half of overdue payments to state and other creditors (December 2013). The insolvency mechanism has yet to prove its effectiveness, as the majority of arrears are over one year old (74 percent, in the case of suppliers, December 2013). The impact of the firms undergoing insolvency proceedings through this indirect channel is notable not only at the time they are declared insolvent but long before that. An important part of the major payment incidents volume (between 20 and 30 percent) is generated by firms under insolvency proceedings one year before entering such resolution mechanism (Graph 9, Annex). Companies that are under insolvency procedure might also have a negative impact on external creditors. These firms are cumulating 2.3 billion euro in debt to non-residents (December 2014), mainly on the medium and long term (58 percent) and only 1 billion euro from their non-resident parent companies. A possible solution to strengthen payment discipline in the economy is to create publicly available databases with the companies experiencing severe financial distress or to improve access to the already available databases on this type of data. At European level, there is an increasing concern regarding enhancing the access to information on firms undergoing insolvency procedures. In July 2014, EU launched a platform for EU wide interconnection of national insolvency registers, currently with only 7 countries, including Romania. The reorganization framework is significantly less used. Only a small number of the companies that have undergone insolvency proceedings are reorganized (4 percent, March 2015). These companies have a relatively larger size (calculated as total assets) and a lower degree of indebtedness compared to the other companies under insolvency procedure. The median value of the debt-to-total-assets indicator is 1.1, while for the rest of the companies is 1.4 (according to data for 2013).

2.2. The Impact of the Insolvency Phenomenon on the Banking Sector Loans granted by banks and NBFIs to companies that entered insolvency procedure form an important share of financial institutions’ portfolio (17.6 billion lei, representing 15 percent of credit to non-financial corporations, March 2015) and are the main contributors to the stock of nonperforming loans (71 percent of nonperforming loans granted to nonfinancial companies).

Central Bank Journal of Law and Finance, No. 1/2015 113 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

However, the insolvency phenomenon is only partially triggered by the debtors’ need to protect themselves from financial creditors as only a fraction of the companies entering insolvency have loans from banks and NBFIs (13 percent during 2014 and the first quarter of 2015). Even when the number of companies with bank credit that have filed for insolvency increased sharply, immediately after the crisis, the number remained contained (Graph 10, Annex). Most of the firms that have undergone insolvency proceedings have arrears towards the state, but this number is on a downward trend (from around 60 percent during 2008-2011, to 48 percent in 2014 and 45 percent in first quarter of 2015). The companies with payments overdue towards suppliers for more than 90 days represent, in average, around 30 percent of the companies entering under insolvency procedure. Moreover, insolvency usually appears at a later stage, when the company is already struggling to meet its payments, including those towards banks. Two thirds of the companies under insolvency procedure present in the banks’ books had already past due payments for more than 90 days at the moment they entered insolvency. Given that banks only recently (June 2014) started to intensify their efforts in cleaning up their balance sheets, the stock of loans granted to companies under insolvency procedure by banks and NBFIs reached a significant level. These loans are, in average, in the nonperforming status for 3 years (the maximum being 7 years). By the type of the collateral associated with these loans, the real estate is the most frequently used (around 85 percent of the loans have, inter alia, a real estate collateral). Only 8 percent of the loans granted to companies in insolvency have no collateral (March 2015). Financial institutions have resorted quite extensively to restructuring measures in the case of loans to companies under insolvency proceedings. Around 30 percent of the loans granted to these companies were rescheduled or restructured, while only a very small amount (2 percent) were refinanced (March 2015). These methods proved not to be too effective in helping the debtor improve its ability to repay the debt but only to delay the moment when the loss will be recognized by the bank. Another set of measures consists of writedowns/ writeoffs and selling. Around one fifth of their credit portfolio was recognized by banks as totally or partially unrecoverable and moved them to off balance sheet, while another 10 percent were sold with an around 70 percent loss. The efficiency of the insolvency mechanism plays an important role in the development of the credit market. By increasing its efficiency, it will positively contribute to the resumption of credit, especially towards the non-financial companies sector as it should decrease the costs incurred by the creditor after the debtor’s default.

114 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

2.3. The Changes to the Insolvency Code The recent changes made to the insolvency legal framework in 2014 should help strengthen the payment discipline, reduce the duration and, implicitly, the costs generated by bankruptcy. The impact on banks' balance sheets will be, most probably, moderate in the near term, given that the new insolvency code does not apply to the outstanding companies under insolvency proceedings. Firstly, by establishing a threshold for the debtor at 40,000 lei at which the insolvency proceedings can be initiated and by reducing the threshold for creditors at the same amount, the new insolvency law ensures an equitable treatment for creditors and debtors and promotes a more prudent behavior among debtors. It might also contribute to a reduction in insolvency applications by debtors and may diminish the abusive use of this resolution mechanism by financial distressed debtors (Pelinescu, 2013). Around 10 to 20 percent of the companies that have undergone insolvency proceedings during 2008 - 2014 had less than 40,000 lei in debt according to their financial statements. Another change introduced by the new code was the reduction from 90 days to 60 days of the number of days overdue above which the company is eligible for filing for insolvency. The measure should strengthen the payment discipline in the economy, but, knowing that the majority of payments due towards suppliers is older than 90 days (91 percent of the overdue payments for the outstanding companies under insolvency procedure at 2015 Q1), additional measures are needed to reduce the amount of arrears in the economy. Finally, the new insolvency law standardizes and provides a much more detailed description for each of the stages of the insolvency procedure, which should help increase the speed, the transparency and, therefore, the predictability of the process. These improvements to the insolvency framework might also contribute to the development of the market for the distress debt.

3. THE EFFICIENCY OF THE INSOLVENCY PROCESS

The efficiency of the insolvency framework can be evaluated both as ex-ante and ex-post. The ex-ante efficiency usually refers to incentives that affect the debtors’ and the creditors’ behaviour, while the ex-post deals with maximizing the value of the firm if the firm is reorganized or minimizing the loss for creditors if the company is liquidated.

3.1. Ex-ante Efficiency To assess the ex-ante efficiency, we dwell upon two different aspects. First, we infer the usefulness of the insolvency mechanism in closing unviable businesses and in reducing the amount of bad debt in the economy by looking at incentives to start the insolvency procedure and the structure of payments due. From this point of view, the assessment of

Central Bank Journal of Law and Finance, No. 1/2015 115 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania the Romanian insolvency framework indicates a relative poor efficiency despite having one of the highest application rates among EU countries. The main reasons are the following: a) a large share of companies in the economy have negative capital (technical insolvency); this might also indicate a high cost of formal bankruptcy and/or voluntary liquidation of the firm, b) the companies undergoing insolvency are usually experiencing financial distress long before entering this resolution mechanism, and c) the largest share of payments due in the economy are older than 90 days. Secondly, we turn our attention to the banks’ credit decisions, as the insolvency framework is a critical factor for the credit market (Hagan, 2015). The existing literature shows that inefficient corporate insolvency laws are usually linked with a higher degree of bank credit with real estate collateral. In Romania, a considerable part of the banks’ credit to nonfinancial companies is backed by a real estate collateral (85 percent in the case of credit granted to the companies under insolvency procedure). Moreover, during the recent economic downturn, the banks increased their requirements regarding collateral, similar to what was observed in other EU countries (Graph 11, Annex). This approach has not helped banks reduce the credit risk of their portfolios. The nonperforming ratio of loans granted to nonfinancial companies backed by a real estate property was higher compared to the sector average (the differences varied between 3 and 10 percentage points during 2010-2014, Graph 12, Annex). The extensive use of the insolvency framework given its limitations and its costs for creditors, coupled with relatively low efficiency of the judicial system, should spur creditors to increase the intensity their monitoring of the debtors’ credit risk. The recent improvement of the Central Credit Register by the National Bank of Romania (NBR) should help banks enhance their assessment of companies’ credit risk. This initiative of the NBR is in line with actions at EU level. The European System of Central Banks (ESCB) has been exploring the potential statistical use of Central Credit Registers (CCRs) from 2007 onwards. The establishment of an IT solution (Analytical System on Credit – AnaCredit) was discussed for receiving, storing and disseminating credit and credit risk information on an EU-wide scale, which would be sourced from national CCRs or other similar datasets. This database would include the most important attributes on loans, lenders and borrowers. In 2014, the ECB decided that the AnaCredit project should have top priority with regard to the upcoming ECB supervisory task. As a consequence, the ESCB must ensure the transfer of European-wide and harmonized credit data to the ECB by the end of 2016. Though the concrete data request has not yet finalized, it is already evident that banks shall report loan portfolio and borrower data on a loan-by-loan level. The access to these databases will allow financial institutions to develop a better capacity to analyze the opportunity of granting credit to viable companies in the economy.

116 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

3.2. Ex-post Efficiency We evaluate ex-post efficiency by analysing the following elements: (i) the duration of the insolvency process and (ii) the loss and/or gains in the firm’s value during reorganization. The duration is an important driving factor of the insolvency cost for the economy. The higher is the time spent in insolvency, the higher are the costs for creditors and lower the probability of reorganization. According to data collected by the World Bank for the Doing Business statistics, around 20 percent of the claims value is lost after one year spent in insolvency. Romania is among the countries with the highest duration of the insolvency process (lower only compared to the Slovak Republic and Greece, and similar to Bulgaria). The average duration of the insolvency process for the companies that were liquidated in 2014 or the first quarter of 2015 was around two years (Table 1, Annex), with important differences between economic sectors. In the industry and agriculture sectors the average duration is higher, at 25 months and, respectively at 26 months, compared to an average of 20 months for the other sectors. Although, the new insolvency law should reduce the time the companies are spending in insolvency, the still large number of firms under the old law (the number of firms under insolvency proceedings is 48.5 thousand firms compared to 20.6 thousand new firms that filed for insolvency in 2014, Table 1, Annex) indicates that the speed of cleaning up unviable businesses and of reducing the level of bad debt from the economy will remain low. The second element that we are interested in this section is the efficiency of the insolvency framework in restructuring debt. A company that enters insolvency has three main exit solutions: (i) out-of-court settlements, if the creditors reach an agreement with the debtor for resolving the financial issue, (ii) bankruptcy, when the firms’ assets are sold and the creditors receive what they are owed or part of that, and (iii) judicial reorganization, when the creditors approve the company’s plan to reorganize its activity and to pay its debts according to a firm schedule. The main goal of the insolvency procedure is to keep viable businesses operating and, therefore, increase the social outcome by preserving the firms’ value and the majority of their employees. On one hand, the companies facing financial distress are interested in reorganizing their activities and restructuring their debt. On the other hand, the creditors want a quick and effective solution, which is the liquidation of the firm. By allowing to correctly pursue the reorganization venue, the insolvency framework proves its efficiency through the additional value created by the firms undergoing reorganization proceedings compared to the value recovered by creditors if the same firms are liquidated. Among EU countries, there is a wide variety regarding the possibility for a firm to ask for reorganization and the way this procedure is conducted. Recently, the Czech Republic

Central Bank Journal of Law and Finance, No. 1/2015 117 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania amended its insolvency law to allow companies to exit the insolvency procedure by way of reorganization. Even when the legal framework allows for such a procedure, it is rarely used due to difficulties in meeting the requirements, including that of reaching an agreement with all the relevant creditors. In Romania, firms that have undergone a reorganization procedure have lost, in average, an important part of their value before or during the year when they begin the reorganization process and only slowly recover afterwards. These firms have usually around 10 percent less assets, 40 percent less employees, 40 percent less market share and creates around 50 percent less gross value added during the year when the reorganization procedure starts. After one year, the firms recover only around 10 percent of the workforce and of their market share compared to the values recorded in the year before entering insolvency (Table 3, Annex).

4. CONCLUSIONS

In Romania, the corporate insolvency is an important process in Romania. Compared to other EU states, Romania has one of the highest utilization rates of the insolvency procedure even compared to countries with similar legal origin. This phenomenon is not backed by a high degree of efficiency. According to the World Bank, Romania is in the last 10 EU countries in what concerns the easiness of resolving insolvency. The companies under insolvency proceedings are mostly companies with Romanian ownership, established during the expansion phase of the economic cycle, are mostly micro companies and are mainly from service and trade sectors. Triggering the insolvency procedure happens at a later stage of the financial distress period, as the companies are usually struggling long before entering insolvency. The direct impact on the economy is relatively limited, but the indirect one, through the payment discipline, is notable. Loans granted by banks and NBFIs to companies that undergo insolvency proceedings form a large share of the financial institutions’ portfolio and are the main contributors to the stock of nonperforming loans. However, the insolvency phenomenon is only partially triggered by the debtors’ need to protect themselves from financial creditors as only a fraction of the companies entering insolvency have loans from banks and NBFIs. Most of them have arrears towards the state, and, to a lesser extent, towards suppliers. Financial institutions have resorted quite extensively to rescheduling or restructuring measures on the loans granted to companies in insolvency. These methods proved not to be too effective in helping the debtor improve its ability to repay the debt but only to delay the moment when the loss will be recognized by the bank. Another set of measures used

118 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța by banks consists of writedowns/ writeoffs and selling, but the banks have only recently started to use them. The financial crisis tested the corporate insolvency mechanism, urging reforms. The changes made to the Romanian insolvency legal framework in 2014 should help strengthen the payment discipline, reduce the duration and, implicitly, the costs generated by bankruptcy. The impact on banks' balance sheets will be, most probably, moderate in the near term, given that the new insolvency law does not apply to the outstanding companies under insolvency proceedings. These improvements to the insolvency framework, by increasing the transparency of the process and the predictability of the outcome might also contribute to the development of the market for distress debt. The Romanian insolvency framework has a relative weak efficiency both as regards ex- ante incentives that affect the debtors’ and the creditors’ behavior, as well as ex-post costs incurred by creditors and loss of the firm’s value when judicial reorganization is chosen. The main reasons are the following: a) a large share of companies in the economy have negative capital (in technical insolvency); this might also indicate a high cost of formal bankruptcy and/or voluntary liquidation of the firm, b) the companies undergoing insolvency are usually experiencing financial distress long before entering this resolution mechanism, c) the largest share of payments due in the economy are older than 90 days, d) the duration of the insolvency process is relatively long, especially for the companies from industry and agriculture sectors, and e) the number of applications for judicial reorganization or out-of-court settlements is low, although increasing, but the firms undergoing reorganization proceedings loose an important part of their value before or during the year they start operating according to their judicial reorganization plan and recover only slowly after that. All these factors might have contributed in shaping the banks’ credit decisions towards a higher preference for real estate assets as collateral when crediting firms or even towards a higher preference for loans to the household sector at the expense of non-financial companies. There is a common understanding that having a well-designed corporate insolvency legal and regulatory framework is not enough. The quality of the judicial system is also a relevant factor in order to have an effective and efficient resolution mechanism and should encourage less risky behavior in the economy by lowering moral hazard. In conclusion, the major results that we find are: (a) the indirect impact on the economy, through the payment discipline is elevated, given the low efficiency of the system to start and to resolve insolvency in the past, and (b) the pressure on the banks’ balance-sheet is high given the stock of nonperforming loans generated by these firms, the majority of which remained under the old insolvency law.

Central Bank Journal of Law and Finance, No. 1/2015 119 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

The main lessons for policy makers that we infer from these results are: (i) enhancing the transparency of the insolvency process, as well as the access to information on the companies experiencing severe financial distress should strengthen payment discipline in the economy; (ii) improving the predictability and the speed of the process, through a good judicial system together with a well-designed legal framework, is essential for achieving an efficient and effective insolvency process and should help the development of the credit market, including the market for the distress debt; and (iii) encouraging creditors to intensify their monitoring of the debtors’ credit risk and not to rely extensively on real estate collateral, should help reduce the risks from their portfolios, including lowering the probability of crediting firms that will eventually enter under insolvency procedure.

120 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

ANNEX - STATISTICS ON CORPORATE INSOLVENCIES

Graph 1: The Number and The Density* Of Corporate Insolvencies in Some European Countries in 2012

Note: The bars with pattern represent countries with French civil law. Data of countries legal origin is from Djankov et al. (2008). * The density term is usually used new businesses density, measured as new registrations per 1000 people ages 15-64 years. We extend this term to corporate insolvencies. The density indicator employed here is calculated as the number of corporate insolvencies per 1000 persons with age between 15-64 years.

Sources: Creditreform Economic Research Unit, Eurostat, Djankov et al. (2008)

Central Bank Journal of Law and Finance, No. 1/2015 121 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

Graph 2: The Ease of Resolving Insolvency, 2014

Note: the bars represent the countries’ rank of easiness of resolving insolvency index. The lower the rank, the better is the insolvency framework. The indicator includes the time, cost, outcome of insolvency proceedings (involving domestic entities) and strength of insolvency framework index (evaluating the adequacy and integrity of the legal framework applicable to liquidation and reorganization proceedings). The data for the resolving insolvency indicators are derived from questionnaire responses by local insolvency practitioners and verified through a study of laws and regulations as well as public information on bankruptcy systems. The ranking of economies on the ease of resolving insolvency is determined by sorting their distance to frontier scores for resolving insolvency. These scores are the simple average of the distance to frontier scores for the recovery rate and the strength of insolvency framework index. Source: World Bank Group, 2015

122 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

Graph 3: Number of Corporate Graph 4: The Improvements in the Insolvency Dynamics between 2008- Insolvency Framework, 2010 - 2015 2012 in Some of EU Countries

Note: the points represent the distance frontier. An economy’s distance to frontier is indicated on a scale from 0 to 100, where 0 represents the lowest performance and 100 the frontier. A score of 80 in 2013 would indicate the economy is improving. Source: Creditreform Economic Research Unit 2012 Source: World Bank Group, 2015

Table 1: Insolvency Statistics and Firm Demographics in Romania

2009 2010 2011 2012 2013 2014 1. Insolvency statistics Number of new firms under 17,776 23,961 22,964 26,759 29,317 20,608 insolvency procedure Density of new firms under insolvency procedure (number of firms under 1.3 1.7 1.7 2.0 2.2 1.5 insolvency procedure per 1000 people ages 15-64 years) Number of firms under 29,928 39,870 45,336 49,278 53,368 48,534 insolvency procedure

Central Bank Journal of Law and Finance, No. 1/2015 123 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

2009 2010 2011 2012 2013 2014 Duration of insolvency proceedings (number of n.a. 13.9 15.1 16.6 19.4 20.9 months, average values)* 2. Bankruptcy statistics Number of firms under 12,951 20,458 24,669 29,590 34,114 31,630 bankruptcy procedure 3. Judicial reorganization and out-of-court settlements Number of firms under judicial reorganization 107 388 762 844 1,052 1,109 procedure Number of firms that had their insolvency procedure 289 437 270 309 339 364 canceled (out-of-court settlement) 4. Firms’ demographics Number of new firms created 56,698 52,041 62,735 61,542 60,292 56,381 in the economy Density of newly created firms (number of new firms per 4.1 3.8 4.6 4.5 4.4 4.2 1,000 people ages 15-64 years) Number of liquidated firms (voluntary or involuntary) in 43,713 43,852** 73,244 71,746 80,786 76,483 the economy Density of liquidated firms (number of liquidated firms 3.1 3.2 5.3 5.2 5.9 5.6 per 1,000 people ages 15-64 years) Number of liquidated firms 7,127 14,317 17,178 23,148 25,116 24,838 after declaring bankruptcy Share of liquidated firms after declaring bankruptcy of total 16.3 32.6 23.5 32.3 31.1 32.5 number of liquidated firms

Note:* calculated as firms that exit insolvency proceedings in the specific year; ** during 2010 an additional 127,294 firms were liquidated as requested by Government Emergency Ordinance 44/2008. We excluded these firms from our analysis. Sources: National Trade Registry Office, Eurostat, authors calculations

124 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

Graph 5: The Structure of Romanian Graph 6: The Financial Stance of the Companies by Economic Sectors at the Romanian Companies under Insolvency Time they enter under Insolvency Procedure in the year before Entering Proceedings, March 2015 Insolvency Procedure

Note: Companies from the utilities sector are less than 10 percent of the services and utilities sector. Data on number of months are calculated only for companies that exit insolvency procedure in 2014 or 2015/Q1

Sources: National Trade Registry Office, Sources: National Trade Registry Office, Ministry of Public Finance Ministry of Public Finance

Central Bank Journal of Law and Finance, No. 1/2015 125 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

Table 2: The Romanian Non-financial Companies’ Assessment on Access to Finance Companies which enter insolvency Total no. of companies (% of procedure after survey date (% of companies which rated the companies which rated the problem as pressing) problem as pressing) Sept. Mar. Sep. Sept. Mar. Sept. Average Average 13 14 14 13 14 14 1. Question on most pressing problems for non-financial companies* Finding 42 47 53 47 45 46 40 44 customers Competition 49 50 46 48 46 47 42 45 Access to 32 38 32 34 20 19 18 19 finance Costs (of production, 44 50 39 44 41 40 37 39 etc.) Availability of 24 23 26 24 25 23 25 24 skilled staff Regulations 29 33 20 27 29 25 21 25 Insolvency 32 33 40 35 19 17 15 17 Payment 52 51 58 54 39 37 34 37 discipline Level of 78 82 78 79 82 81 73 79 taxation 2. Question on sources of financing accessed by non-financial companies, 6 months prior to survey date Retained 34 32 26 31 43 44 46 44 earnings or sale of assets Loans from 28 31 25 28 29 31 35 32 shareholders or capital increases Bank overdraft, 23 24 25 24 13 12 8 11 credit lines or credit cards overdraft Cash flow loan 7 6 7 7 3 2 1 2 Investment loan 9 7 6 7 5 4 4 4 Trade credit 37 33 29 33 21 19 14 18 Financial 12 12 5 10 7 6 4 6 leasing or factoring

126 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

Companies which enter insolvency Total no. of companies (% of procedure after survey date (% of companies which rated the companies which rated the problem as pressing) problem as pressing) Sept. Mar. Sep. Sept. Mar. Sept. Average Average 13 14 14 13 14 14 3. Question on reasons for which the company accessed financing, 6 months prior to survey date** Bank debt 7 7 7 2 2 2 refinancing Working capital 48 41 45 35 34 34 or payment to suppliers Investments in 12 9 10 9 8 9 business development You have not 33 42 38 53 56 55 accessed external financing over the past 6 months Note * * question with grading scale (from 1 – highly pressing to 5- not pressing), where only grades 1 and 2 were considered to indicate problems considered pressing. ** changes were performed after September 2013 and the results from that date are not comparable with the rest. Source: National Bank of Romania, Survey on the access to finance of non-financial corporations in Romania and their capacity to withstand adverse financial conditions.

Graph 7: Cumulative Distribution of the Graph 8:Cumulative Distribution of Debt to Equity Indicator for Romanian the Liquidity Ratio for Romanian Companies, December 2013 Companies, December 2013

Source: National Trade Registry Office, Ministry Source: National Trade Registry Office, of Public Finance, authors calculations Ministry of Public Finance, authors calculations

Central Bank Journal of Law and Finance, No. 1/2015 127 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

Graph 9: Major Payment Incidents Graph 10: The Importance of Bank Credit Generated by Romanian Companies for the Romanian Companies under under Insolvency Procedure by Year of Insolvency at the Time they Enter Entering under Insolvency Procedure Insolvency Procedure (% of total major payment incidents generated in the economy in a specific year)

Note: t represents the year when insolvency Note: The bars represents the volume of credit proceedings were initiated. For example, in granted by banks at the time the companies 2014 the major payment incident generated by enters under insolvency proceedings (on the left firms that underwent insolvency proceedings hand side), while the line represents the were 38.5% (from which 30% in the same year percentage of companies under insolvency they enter under insolvency procedure, 4.3% procedure with loans from banks at the time by firms that enter insolvency in 2013, they are declared insolvent (on the right hand displayed in the t+1 section as 2013 bar, and side). 3.5% by firms that enter insolvency in 2012, showed in the graph in the t+2 section as 2012 bar). Source: National Trade Registry Office, Source: National Trade Registry Office, National Payment Incidents Register, authors Bank of Romania, authors calculations calculations

128 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

Graph 11: Change in the Collateral Graph 12: Nonperforming Ratio of Loans Requirements by Banks Granted to Nonfinancial Companies by (net percentage) Romanian Banks

Note: Higher values represents a tightening of the collateral requirements

Source: European Central Bank, National Bank of Romania, Bank Lending Survey Source: National Bank of Romania

Table 3: Statistics of the Romanian Nonfinancial Companies undergoing Judicial Reorganizational Procedure

Values relative to the year before entering under insolvency procedure

Number of Gross value Assets Turnover employees added t t+1 t t+1 t t+1 t t+1 All firms Mean 1.0 1.0 0.7 0.7 0.5 1.1 1.0 1.1 Median 0.9 0.9 0.6 0.7 0.5 0.6 0.6 0.7 Standard 1.0 0.8 0.7 0.6 4.8 8.0 2.2 4.6 deviation

Central Bank Journal of Law and Finance, No. 1/2015 129 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

Values relative to the year before entering under insolvency procedure

Number of Gross value Assets Turnover employees added t t+1 t t+1 t t+1 t t+1 Corporations Mean 0.7 0.7 0.4 0.5 0.3 0.2 0.5 0.6 Median 0.7 0.8 0.4 0.6 0.4 0.5 0.5 0.6 Standard 0.4 0.3 0.3 0.3 1.2 2.0 0.4 0.4 deviation

Note: t represents the year when the companies enters under the reorganization procedure Source: National Trade Registry Office, National Bank of Romania, Ministry of Public Finance

130 Central Bank Journal of Law and Finance, No. 1/2015

Irina Mihai, Alina Tarța

REFERENCES

1. Biais, B.; Mariotti, T., 2009, Credit, wages, and bankruptcy laws. Journal of the European Economic Association, 7(5), 939-973 2. Claessens, S.; Klapper, L. F., 2005, Bankruptcy around the world: Explanations of its relative use. American Law and Economics Review, 7(1), 253-283 3. Creditreform Economic Research Unit, 2012, Corporate insolvency in Europe, http://www.creditreform.de/index.html 4. Dahiya, S.; Klapper, L., 2007, Who survives? A cross-country comparison. Journal of Financial Stability, 3(3), 261-278 5. Davydenko, S. A.; Franks, J. R., 2008, Do bankruptcy codes matter? A study of defaults in France, Germany, and the UK. The Journal of Finance, 63(2), 565-608 6. Djankov, Simeon; Edward Glaeser; Rafael La Porta; Florencio Lopez-de-Silanes; Andrei Shleifer, 2003, The New Comparative Economics, Journal of Comparative Economics 31, 595-619 7. Djankov, S., McLiesh, C.; Shleifer, A., 2007, Private credit in 129 countries. Journal of financial Economics, 84(2), 299-329 8. Djankov, Simeon; Oliver Hart; Caralee McLiesh; Andrei Shleifer, 2008, “Debt Enforcement around the World.” Journal of Political Economy, 116(6): 1105–49. 9. García-Posada, M. (2013). Insolvency institutions and efficiency: the Spanish case (No. 1302). Banco de Espana 10. Hagan, S. (2000). Promoting Orderly and Effective Insolvency Procedures. Finance and Development (Volume 37, No. 1), International Monetary Fund 11. Hagan, S. (2015). Bankruptcy Conference: The World of Insolvency, Boston College Law School 12. Hussain, Q.; Wihlborg, C., 1999, Corporate Insolvency Procedures and Bank Behavior-A Study of Selected Asian Economies (No. 99-135). International Monetary Fund 13. Jappelli, Tullio; Marco Pagano, 2002, Information Sharing, Lending, and Defaults: Crosscountry Evidence, Journal of Banking and Finance 26, 2017-2045 14. Mulligan, Casey; Andrei Shleifer, 2005, The Extent of the Market and the Supply of Regulation, Quarterly Journal of Economics 120, 1445-1474 15. National Bank of Romania, Survey on the access to finance of non-financial corporations in Romania and their capacity to withstand adverse financial conditions, http://www.bnro.ro/Regular-publications-2504.aspx 16. National Bank of Romania, Bank Lending Survey, http://www.bnro.ro/Regular- publications-2504.aspx

Central Bank Journal of Law and Finance, No. 1/2015 131 The Role of the Insolvency Framework in Strengthening the Payment Discipline and in Developing the Credit Market in Romania

17. Padilla, A. J.; Requejo, A., 2000, The Cost and Benefits of the Strict Protection of Creditor Rights: Theory and Evidence (No. 3084). Inter-American Development Bank, Research Department 18. Pelinescu, E., 2013, The mechanism of arrears in Romania, Romanian Journal of Economic Forecasting, 2, 223-239 19. World Bank Group, 2015, Doing Business 2015, www.doingbusiness.org

132 Central Bank Journal of Law and Finance, No. 1/2015

The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective1

Dan Pălăngean*

Abstract

The paper analyses the evolution of Romania’s GDP after 1989, in relation with the funds assigned to work (compensation of employees), part of which is used for the payment of past work (pensions). The transition from socialism to capitalism did not have important consequences for the purchasing power of employees and pensioners, due to the primitive accumulation of capital. The paper draws several conclusions based on the evolution of the labour share in GDP, costs of changing the social system, mobility of workforce abroad and implications of the transition to a single European currency.

Keywords GDP, wage income, pension, purchasing power, workforce, budget deficit

JEL Classification: E01, H62, J30, J26, J61, J11

1 Fragments of this paper can be found on Contributors.ro signed with the pseudonym Marin Pană, which is, in fact, an homage to my grandfather. * Strategy Adviser, National Bank of Romania.

Central Bank Journal of Law and Finance, No. 1/2015 133 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

1. EVOLUTION OF THE GDP OF ROMANIA IN THE PAST 25 YEARS

25 years after December 1989, statistics show an economic growth of our country as if it had only covered a Five Year Plan like in the old days. However, this time statistics do not show a centralized economy and doubtful reports, but the evolution of a fully-fledged EU member, integrated in what used to be called “consumer society”. Before the regime shift, Romania had reached the maximum level of development measured in the GDP from 1988, when it slightly exceeded the 60-billion-dollar threshold. The most favoured nation status, still in place in the relationship with the USA, was renounced; then, in the spring of 1989, it was announced that the external debt was paid in full. Unfortunately, the pressure of the massive annual payments made in order to eliminate external debt and the lack of investments in advanced technologies had exhausted the economy. The decline came next, even before the collapse of Ceaușescu’s regime. Therefore, in 1990 we only attained two thirds of the maximum value of the socialist GDP in USD terms. Current values in lei were becoming irrelevant because of the hyperinflation. Apart from inherent drawbacks, this allowed the fast adjustment of the ratios between the relative prices of goods and services.

Table 1: Evolution of GDP in 1985-1989

Year 1985 1986 1987 1988 1989 1990 GDP (bill. lei) 817 838 845 857 800 858 GDP (bill. USD) 47.6 52 58 60 53.6 40.8 Source: NIS, data recalculated based on the European System of Accounts (ESA) 2010

The decrease in production, the disbandment of the Council for Mutual Economic Assistance (CMEA) commercial system of the socialist countries and the loss of marketplaces, along with the liberalisation of prices and the exchange rate led us, after two years, to the catastrophic situation in which we went down to less than a third of the value in dollars stated in 1988. Then, a sinuous comeback came next, marked by the world crisis and by a negative economic growth at the end of the decade.

Table 2: Evolution of GDP in 1991-1999

Year 1991 1992 1993 1994 1995 1996 1997 1998 1999 GDP (thousand bill. lei - 2.2 6.0 20.0 49.7 72.1 109 252 338 545 ROL) GDP (bill. USD) 28.9 19.6 26.3 30.0 35.4 35.3 35.2 38.1 35.6 Source: National Institute of Statistics (NIS), data recalculated based on the ESA 2010

134 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

The grounds of the attempt to boost production in 1992-1996 proved to be unsolid. In 1997-1999, we dealt with big macroeconomic problems and the Romanian state was within an inch of payment cessation. The problem is that we missed out on about a decade of development and it was only in 2000 that we came back to the performance from 1991. It is only since then, after shifting the focus point on the private sector and on reforming the company taxation system, that the economic growth has truly started. The recommencement of the economic growth had only picked up the pace at some point after 2000 and the advance had been ongoing until 2008, when the GDP reached the greatest historical value in real terms. Table 3: Evolution of GDP in 2000-2014

Year GDP (bill. lei – RON) GDP (bill. euros) 2000 80.4 44.8 2001 116.8 48.4 2002 152.0 50.3 2003 197.4 58.9 2004 247.4 61.1 2005 289.0 79.8 2006 344.6 97.8 2007 404.7 121.3 2008 514.7 139.8 2009 501.1 118.3 2010 522.6 124.1 2011 557.3 131.5

2012 596.7 133.9 2013 639.3 144.7 2014 666.6 150.0

Source: NIS, data recalculated based on the ESA 2010

Central Bank Journal of Law and Finance, No. 1/2015 135 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

After a quarter-century since the regime shift from 1989, Romania is about to resume the economic growth from where it was left in 2008, given the almost full recovery of the decline from 2009-2010. This time, all the macroeconomic indicators are in accordance with the requirements imposed by the Maastricht Treaty, with a stabilised inflation and an exceptionally stable exchange rate. The question is whether we are able to reach and maintain a growth rhythm in a medium term (five years), which would allow us to close the gap to the EU average in terms of GDP per capita with more than 2 percent annually, in order to exceed the recommended threshold of 60 percent of this average and to join the Euro on acceptable terms.

2. THE CONTRIBUTION OF THE COUNTIES TO THE GDP SHOWS ENORMOUS DEVELOPMENT DISPARITIES

The data presented by the National Commission of Prognosis regarding the contribution of each county to the GDP nation-wide show a significant disparity between the capital region and other development regions, and especially within these regions. Bucharest represents more than a quarter of Romania’s economy, while the least developed 11 counties bring less than 10 percent of the GDP. To put it more simply, the situation looks like a combination of seven development regions that bring about 10 percent of the GDP on average plus the capital region, which brings almost all remaining 30 percent. South Muntenia is slightly above the average and South-West Oltenia is significantly below the average (see Table 4).

Table 4: GDP per Regions

Region GDP (bill. lei) Share in GDP (%) North-East 67.4 10.0 South-East 67.9 10.0 South Muntenia 82.4 12.2 South-West Oltenia 55.0 8.2 West 68.1 10.1 North-West 70.3 10.4 Centre 74.7 11.1 Bucharest-Ilfov 188.0 27.9 Total 674.3 100 Source: NIS, personal calculations based on NIS data Given the higher population density in the Capital area, the division of the actual development regions from an economic standpoint seems to be quite balanced. The major

136 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean differences are not between these regions, but within them, where we notice a relatively strong county, at a ratio ranging from 3.5:1 to 5:1 in comparison with the weakest county (see Table 4a).

Table 4a: Ratio of GDP inside Regions

City Ratio North-East Region 3.5:1 Bacău 20% Botoşani 10% Iaşi 29% Neamţ 13% Suceava 20% Vaslui 8% South-East 5:1 Brăila 11% Buzău 13% Constanţa 39% Galaţi 19% Tulcea 8% Vrancea 10% South Muntenia Region 4:1 Argeş 24% Călăraşi 8% Dâmboviţa 14% Giurgiu 8% Ialomiţa 8% Prahova 31% Teleorman 8% South-West Oltenia 4.5:1 Dolj 41% Gorj 20% Mehedinţi 9% Olt 15% Vâlcea 16%

Central Bank Journal of Law and Finance, No. 1/2015 137 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

City Ratio North-West 5:1 Bihor 22% Bistriţa-Năsăud 10% Cluj 34% Maramureş 15% Satu-Mare 11% Sălaj 7% Centre Region 5:1 Alba 15% Braşov 30% Covasna 6% Harghita 9% Mureş 19% Sibiu 21% West Region 4:1 Arad 23% Caraş-Severin 12% Hunedoara 17% Timiş 49% Bucharest-Ilfov Region Ratio irrelevant for this particular region Ilfov 9% Bucharest 91% Source: Personal calculations based on NIS data The first and the second county in terms of economic force succeed in almost all cases in having the majority share in the registered results; the regionally formed tandems are Constanța-Galați (accounts for 58 percent of the GDP of the region they belong to), Prahova-Argeș (55 percent), Dolj-Gorj (61 percent), Timiș-Arad (72 percent) and Cluj- Bihor (56 percent). The North-East and Centre regions are alike as far as the inner distribution is concerned because they are at approximately 50 percent with the first two counties and 70 percent with the first three, given that the “first runner up” position is doubled. Iași is followed by Bacău and Suceava, with almost identical shares and Brașov is followed by Sibiu and Mureș, with a very small difference between them.

138 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

The intraregional GDP distribution that looks like a leopard’s fur (a few component counties intensely coloured by results and the majority comprising the counties left behind) is nationally coherent. Each development region has a representative in the nation-wide top of shares in GDP, where we drew the line at 3 percent (or almost 3 percent).

Table 5: Top Counties with the Highest Contribution to GDP

County Region Share in GDP 1. Timiş West 4.94% 2. Constanţa South-East 3.96% 3. Prahova South 3.76% 4. Cluj North-West 3.57% 5. Dolj South-West 3.30% 6. Braşov Centre 3.28% 7. Argeş South 2.99% 8. Iaşi North-East 2.88% Source: personal calculations based on NIS data South Muntenia is an exception because it has two representatives in the top: Prahova and Argeș, the hard core of the Romanian economy, at a small distance from the capital. The extreme positions in the ranking of the regional representatives, Timisoara – Iași axis, would outline quite well a correlation of decrease in the standard of living from West to East and from North to South, at a quick glance. However, there are two problems concerning poverty concentration in this landscape. The first one is in the centre of the country, where (through the small dimension and rather difficult geographical conditions) Covasna county is by far the least significant in terms of economic results, both nationally and regionally speaking. Its neighbour, Harghita (the other county with a Hungarian majority), is not doing so well either, but it is on the 11th place nationally, slightly surpassing the symbolical threshold of 1 percent to the national GDP. However, it is surprising that the word “South” appears in 6 of the weakest 11 counties (the group of five plus Mehedinți competing with Covasna).

Table 6: Top Counties with the Lowest Contribution to the GDP County Region Share in GDP 1. Covasna Centre 0.70% 2. Mehedinţi South-West 0.74% 3. Sălaj North-West 0.78% 4. Vaslui North-East 0.83% 5. Tulcea South-East 0.83%

Central Bank Journal of Law and Finance, No. 1/2015 139 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

County Region Share in GDP 6. Ialomiţa South 0.92% 7. Giurgiu South 0.92% 8. Vrancea South-East 0.97% 9. Teleorman South 0.98% 10. Botoşani North-East 1.01% 11. Harghita Centre 1.03% Source: personal calculations based on NIS data Consequently, we are dealing with a true poverty belt, spread throughout the South of the country, from Banat to the Delta, which is combined, however, with the relative prosperity of the counties in proximity (including the area of the capital city), in their majority (except for positions 4, 6 and 8 in the top of the high shares) in the lower half of Romania, as it appears on the map. The relative disparities of development nation-wide are perceived here, not in the North-East Region. Over there, poverty as a phenomenon is definitely at a larger scale, but it is less painful because it manifests more uniformly (we also encounter here the minimum disparity between the counties in a given region). This is food for thought, at least when it comes to the funds distribution approaches (including European and irredeemable) meant to ensure the reduction of the development disparities, with the possible change of emphasis from the inter-regional profile to the intra-regional profile.

3. THE EVOLUTION OF THE REAL SALARY AND OF THE REAL PENSION IN RELATION TO THE GDP SINCE 1990

Data from the Statistical Annual 2014 published by the National Institute of Statistics show that the real salary only came back to the level from 1990 when we joined the European Union, in 2007. The minimum level was reached in 1997, when the purchasing power dropped to 56.2 percent of the one registered at the onset of the liberalization of prices. In 2008, pensioners went back to the reference standard of living from 1990, after an exceptional rise in revenues of 65 percent in real terms in only two years. Pensions passed through the minimum purchasing power point in 2000, when they reached 44 percent in real terms compared to the reference year 1990. It can be easily noticed that both minimum values of the purchasing power, calculated for wage or for pension, are well below the minimum reference of the GDP in real terms, namely 75 percent from 1990 reached in 1992. The differences between GDP – wage indices and GDP – pension indices show how space was created for the accumulation of local capital and for the attraction of foreign capital at the level of economic results distribution.

140 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

Table 7: Evolution of the Real wage Index in 1990-2014*

Real Wage Wage-GDP Pension-GDP Year GDP Pension Index** Index Disparity Disparity 1990 100 100 0 100 0 1991 82.2 81.5 -0.7 74.3 -7.9 1992 75 70.8 -4.2 63.1 -11.9 1993 76.1 58.9 -17.2 56.3 -19.8 1994 79.1 59.1 -20 55 -24.1 1995 84.7 66.5 -18.2 61.4 -23.3 1996 87.4 72.7 -14.7 63.1 -24.3 1997 83.1 56.2 -26.9 50.3 -32.8 1998 81.3 58.4 -22.9 49.2 -32.1 1999 81.1 57 -24.1 47.2 -33.9 2000 83 59.4 -23.6 44 -39 2001 87.7 62.4 -25.3 46.6 -41.1 2002 92.2 63.9 -28.3 48.2 -44 2003 97 70.8 -26.2 51.6 -45.4 2004 105 78.3 -26.9 57.7 -47.5 2005 110 89.5 -20.2 62.3 -47.4 2006 118 97.4 -20.9 68.1 -50.2 2007 127 111.8 -14.7 83.6 -42.9

2008 137 130.3 -6.9 112.1 -25.1 2009 128 128.3 0.8 125.7 -1.8 2010 127 123.6 -2.9 122.6 -3.9 2011 128 121.3 -6.6 116.8 -11.1 2012 129 122.5 -6.1 116.9 -11.7 2013 133 123.4 -9.7 117 -16.1 2014 137 127.1 -9.9 120.5 -16.5

Source: NIS, personal calculations based on NIS data

*monthly arithmetical average, **pensioners’ real average pension of state social securities

Central Bank Journal of Law and Finance, No. 1/2015 141 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

The real wage index, obtained by relating the increase in wages to the price rise, expresses the purchasing power in relation to the initial reference point. It dropped instantly, in only three years, under the threshold of 60 percent of the value from 1990 and it stood there for 6 years of the last decade of the last century (1993-1994 and 1997-2000). The comeback attempt from 1996 was suddenly stopped because of the economic imbalances accumulated during the growth without restructuring and it was only in 2004 that the wage level from 1996 was reached again due to the stable growth. The decision made in 2000 to catch, at all costs, the European integration train in 2007 was crucial for the comeback of the real wage to the reference point from 1990. Obviously, the stake set too high in 2008 in relation to the results in economy (see the only slightly positive correlation related to the reference GDP level) could not be kept. The economic crisis was followed by a rearrangement above the threshold of 120 percent and by the gradual recommencement of the comeback, this time on solid grounds, to the purchasing power and the related standard of living registered in 2008. However, for now, we still have to work to get there. The evolution from 2014 of the real wage index shows that the comeback did not succeed, not even to the values from 2009, despite the public sector employee salary recalculation and the advance of the private sector or the historically low levels of inflation (only 1.07 percent annual average and 0.83 percent from December to December for 2014). In simpler words, the situation from 1996-2004 will repeat itself (in another form and in another context) and the clear crossing under conditions of tenability of the threshold established in 2008 will only be achieved in 2016, eight years after the failure of the electively-forced rise of the population’s incomes. Consequently, it would be important not to repeat the cycles of sudden growth followed by a drop with a long comeback and to take a less spectacular, but ongoing advance path.

4. THE EVOLUTION OF THE REAL INCOMES – KEY-FACTOR

Shortly after the events from 1989, the National Commission for Statistics, as NIS was called back then, issued an abstract presenting the main economic indicators of Romania in their non-cosmeticized version for the party plenary sessions. Some evolutions are worth mentioning in order to understand better what generated the popular uprising and what the state of the economy was. The most important indicator, from this perspective, was the real wage (nominal amount received by an employee and adjusted according to price rise). The lowest point of Romanians’ purchasing power during 1981-1990 was reached in 1988, exactly when the maximum level of the GDP from the socialist period was also reached.

142 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

Table 8: Dynamics of the Real Wage (1980 =100)

Year 1985 1986 1987 1988 1989 1990 Real Wage 95.9 95.1 93.4 93.2 95.2 100.5

Source: NIS In other words, the effort of paying the external debt and the limitation of the internal consumption in order to export as much merchandise as possible overlapped the reduction in the standard of living. This happened after “labour people” had got used to wage-indexing to the extent of tens of percentages per five years and to price stability after 1965. The wage rise from 1990 did nothing but to bring the purchasing power again to the level from 1980, after a decade of policy of imposed austerity for the forced payment of the external debt massively contracted after the agreement with the IMF from June 1981. It is worth mentioning that the number of employees contributing to this payment is 70 percent higher than the current one.

Table 9: Number and Share of Employees in the Occupied Population

Year 1980 1985 1988 1989 1990 Employees (thousands) 7,378.5 7,689.3 7,876.8 8,023.8 7,902.4 - of which workers (thousands) 5,891.5 6,104.1 6,264.6 6,408.2 6,167.7

Employees’ share in the occupied 72.6 72.9 73.3 71.3 population 71.3

Employees per thousand 331 338 341 346 341 inhabitants

Source: NIS About one in four Romanians willing to work went abroad because of the lack of jobs and/or poor wage. The result was the loss of an important source for the state budget, from the income tax and social security, to VAT and excise. This explains the budgetary income, placed somewhere around three quarters of the European average of the share in GDP (for a current number of 260 employees per 1 000 inhabitants we obtain exactly this result). By means of the amounts sent back to the country, “the strawberry pickers” (name given to the people who went to work in other European countries, particularly Spain) have become the most important “foreign” investor in Romania, financed the current account deficit and have indirectly supported the exchange rate of the leu in relation to the main foreign currencies. As far as the Romanians left in the country are concerned, the structure of incomes from 2013 was remarkably close to the one registered in 1990 (see Table 10).

Central Bank Journal of Law and Finance, No. 1/2015 143 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Table 10: Structure of Population’s Total Incomes

Year 1980 1985 1988 1989 1990 2013 Work incomes 75.3 77.0 76.1 75.5 74.6 75.4 -from wages 55.9 55.4 55.0 55.9 51.2 51.2 -from agriculture 16.9 19.0 18.7 17.3 20.7 19.8 -other incomes 2.5 2.6 2.4 2.3 2.7 2.6 Social incomes 21.7 20.5 21.5 22.1 22.0 22.8 Other incomes 3.0 2.5 2.4 2.4 3.4 3.4

Source: NIS Data presented by NIS show that the share of wages in the total of incomes stood at over 55 percent during 1980-1989 and dropped towards (sometimes under) 50 percent nowadays. Strangely enough at first sight, social incomes and benefits offered to those in need were actually lower as shares in the total revenue in the socialist period than currently. There are elements that should be taken into account when calibrating the social support policies developing currently, in the sense that the best social protection is ensured by the creation of jobs, instead of indexing the benefits offered from the budget. It is very interesting that self-consumption from one’s own agriculture sources stood at almost 20 percent, which shows that this productive sector has failed to pass towards the developed market economy and that it has missed the opportunity to develop a truly competitive agriculture at European level.

Table 11: Evolution of Real Average Pension (1980 = 100)

Year 1985 1988 1989 1990 State Social Security Pensioners 91 90 94 101

However, the situation of the real incomes started to readjust in 1989. We will not have the opportunity to find out what could have happened after 1990 with the money annually destined to pay the foreign debt, specifically in what projects two to three billion dollars per year (back then the dollar was worth double than it is now in terms of purchasing power) would have gone.

144 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

5. THE STANDARD OF LIVING INCREASED DURING THE CRISIS. THE POSITION AT A EUROPEAN LEVEL FROM THE PERSPECTIVE OF THE ACCESSION TO THE EUROZONE

Ever since the effects of the world economic crisis manifested, Romania has improved its standard of living indicator from 49 percent of the EU average to 55 percent of this average. The advance from this period registered by Eurostat is placing Romania on the third place, after Poland and Lithuania, on the same level with Latvia. It is important to notice that, despite the appearances, data show that, in relation to the European average, the standard of living from Romania never dropped during the economic crisis. The synthetic indicator used for the standard of living was the Gross Domestic Product at standard purchasing power parity. Currently, a Romanian lives averagely at a standard of living of 55 percent of the EU average, with smaller incomes and prices (expressed in Euro) than at the level of the Union.

Table 12: Evolution of GDP/Inhabitant (PPS) in 2008-2013

Country 2008 2009 2010 2011 2012 2013 EU 28 100 100 100 100 100 100 EA 18 108 108 108 108 108 108 Belgium 116 118 120 120 120 119 Bulgaria 43 44 44 44 45 45 Czech Rep. 81 82 81 83 82 82 Denmark 124 123 128 126 125 124 Germany 116 115 120 122 123 122 Estonia 69 64 64 68 71 73 Ireland 131 128 128 130 130 130 Greece 93 95 89 77 74 73 Spain 103 103 99 95 94 94 France 107 109 109 108 107 107 Croatia 65 63 60 60 61 61 Italy 104 104 103 103 101 99 Cyprus 99 100 97 96 93 89 Latvia 58 54 55 57 60 64 Lithuania 64 58 62 65 69 73 Luxembourg 263 252 262 265 264 257 Hungary 64 65 66 65 65 66 Malta 81 84 87 84 84 86

Central Bank Journal of Law and Finance, No. 1/2015 145 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Country 2008 2009 2010 2011 2012 2013 Netherlands 134 132 133 135 132 131 Austria 124 126 126 128 129 128 Poland 56 60 63 64 66 67 Portugal 78 80 80 78 76 79 Romania 49 50 51 51 53 55 Slovenia 91 86 84 83 82 82 Slovakia 72 73 74 73 74 75 Finland 119 114 114 117 115 113 Sweden 124 120 123 127 126 127 Great Britain 114 112 108 106 107 109 Source: Eurostat The evolutions from several European countries allow us to draw some conclusions regarding the relative modifications of the standard of living, according to the adopted economic and social policies. Germany’s performance can be easily noticed, the generator of the economic relaunch, a great economy, which has significantly improved its relative position and has recorded the most significant growth, although it is only on the fourth place. We should also note that, given the position of a second commercial partner of Romania, Italy has also gone below the average standard of living in the EU. Poland, Lithuania and Latvia, as well as Austria, have managed the economic challenges well and, benefitting from Germany’s training effect, have marked a remarkable growth in the standard of living. We should keep in mind that Baltic economies (more elastic from the perspective of the reduced dimensions and helped by its northern neighbours) had a rebound period in which they rapidly restructured the bases on which they started the economic relaunch. In this context, Romania was most similar to Poland, with a gain of four percentage points after 2010. There is hope, but we should also take into account the warning regarding the potential accession to the Eurozone, under the circumstances in which the economy would not be able to deal with the competition in the medium and long term. In the charts of the relative decreases in the standards of living, the apparition of Slovenia is surprising, a country belonging to the Eastern bloc, which had got really close to the target of reaching the 100 percent threshold. Among the former colleagues from the East, we should also underline the underperformance of Croatia, which joined the EU based on a rebound of the standard of living, as well as the stagnation of Hungary. The latter was surpassed by Poland after it had stood at the same level in the last three analysed years.

146 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

Basically, for the next four years, we need a similar advance as in the last four years, if not better, a rise in the GDP per capita to close the gap with the centre of Europe. All these considerations gain special importance through the declared objective of Romania, to join the Eurozone in 2019. Without being an expressed requirement and beyond Hungarians’ exaggerations (which see a necessary of 90 percent of the EU average), the minimal reference point for this accession would be a standard of living of 60 percent of the one registered in the entire Union. An increase of 5 percent in the standard of living expressed through the GDP at the standard purchasing power parity in the next four years remains problematic, therefore the calendar of the restructuring measures needed will be very tight. Within the context, price stability will become at least just as important as GDP growth, so that the inflation would not wear down the economic advance.

6. THE DEFICIT GENERATED BY PENSIONS

A few years before joining the European Union, Romania had a robust development pace, including by reducing costs with social security and by maintaining pensions at a low level. The significant increase in pensions at the end of 2008, when the continuation of the robust economic growth after 2000 was still claimed, inevitably led to a fundamental imbalance and chronic compensation of the social security budget from the state budget. The disparity between the income and expenditures of this budget, mainly dedicated to the payment of pensions, rapidly went up to over 2 percent of the GDP and reached its maximum value related to the results of the economy in 2011. This thing happened, despite the delay of the annual pension indexation, after the Constitutional Court’s decision to maintain the value of the pension point (diminished by 15 percent through the Government’s accountability). For several years, an unfavourable process has been taking place. The balancing (required by law) of the social security budget has needed more and more consistent transfers from the state budget, based on the amounts raised as excises, VAT, income taxation, profit taxation etc., which means two things: first, a great part of the money raised through taxes and dues is sent in consumption, instead of investment, so instead of development, redistribution (only necessary to the segment of reduced incomes) is being performed; the second, beneficiaries of the pensions pay themselves a greater and greater part of the pensions they are benefitting from through taxes and dues (paid by them). Namely, at the distribution of the economic results, they receive in real terms a smaller part than the one on their pension payslip. Last year, the amount transferred from the state budget for the payment of the pensions went down to under two percentages of GDP, but in 2014 it came back above the threshold of 2 percent of GDP, despite a bigger economic growth than the one initially estimated.

Central Bank Journal of Law and Finance, No. 1/2015 147 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Table 13: Evolution of the Transfers from the State Budget to the Social Security Budget (2007-2014)

Year Transferred amount (mill. lei) Share in GDP (%) 2007 11.7 0 2008 1,379.6 0.27 2009 6,397.5 1.26 2010 10,954.7 2.14 2011 13,329.3 2.43 2012 13,148.7 2.24 2013 12,254.3 1.97 2014 13,471.9 2.01 Source: Ministry of Finance, personal calculations In 2010, the Constitutional Court of Romania blocked the adjustment of the pension point value down to an economically tenable value (measure for which the Government had taken responsibility in the Parliament), which led to the increase in VAT with five percentage points. This approach has taken the disinflation process off the trajectory and burdened the business environment, which has also led to the increase in tax evasion. However, was the faster increase in pensions really more necessary than the one in wages? Basically, if we take a look at the evolution of the pension point values, we notice that it doubled in nominal expression from 2007 to 2014. The 5 percent indexation for 2015 led to its increase up to 109.5 percent in nominal terms during the eight years since the accession to the EU. Meanwhile, the price increase was equal to 45.5 percent in the same interval. During the same time span, the increase in wages was at 99 percent in gross value (to which dues fuelling incomes for pensions are also related), but only 92 percent in net value (the significantly different values show that we do not have a flat income tax, but a “discretely progressive” one).

Table 14: Increase in the Pension Point Value and Wage Incomes in 2007-2014

Nominal Real Increase in pensions 99.5% 37% Increase in gross wage incomes 98.7% 37% Increase in net wage incomes 91.9% 32% Increase in prices 45.5% - Source: personal calculations based on NIS data

148 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

If we put together these increases in comparison with 2007, the last year in which Romania did not need to transfer money from the state budget to the social security budget, in order to pay off the pensions, it can be easily seen where the chronic deficit of public income comes from. Under these circumstances, the following question is raised: why should have these social benefits been generally increased by expanding the public cliché of the needs of the elders with low incomes, outside the reality of the amounts collected by public finances, thus affecting the state budget, increasing taxation and hindering economic growth?

Table 15: Evolution of the Pension Point Value in 2007-2015

Month and Year Pension Point Value January 2007 396.20 lei September 2007 416.00 lei November 2007 541.00 lei January 2008 581.30 lei October 2008 697.50 lei April 2009 718.40 lei October 2009 732.80 lei January 2013 762.10 lei January 2014 790.80 lei January 2015 830.20 lei

Basically, given the impossibility of increasing social security contributions (which would impermissibly grow costs that are already high with the taxation of the workforce), a deficit of approximately two percentages of GDP was agreed in the social security budget in the long term. The simple and logical solution, under everybody’s nose, would be for the pension point value not to be set arbitrarily, at politicians’ hand, but to result from the total amount of social contributions of working people, divided by the total number of pension points beneficiaries are entitled to, according to the law; then to ensure an additional social protection only for the persons with low incomes, for this reason and not as a result of previous work. Anyhow, given the origins of the amounts received, the pensioners’ standard of living depends absolutely on the economic performance of the active generation and not relatively (between them, as an internal hierarchy) on the work from years ago.

Central Bank Journal of Law and Finance, No. 1/2015 149 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

7. ROMANIA, ON THE LAST PLACE IN EUROPE IN TERMS OF REMUNERATING PRIVATE LABOUR

The recalculation of the quarterly Gross Domestic Product as of 1995, through incomes based on the new European methodology ESA 2010, has allowed the emphasis on the unitary basis of the wage evolutions from Romania in the last 20 years. The components we selected are employees’ remuneration (gross volume, including social contributions) and the share of those working for the state in the total of remuneration. It can be noticed how employees’ remuneration has varied significantly in Romania ever since 1995 and how the maximum was reached in 2001, the only year in which the threshold of 40 percent was exceeded. In the last two years analysed (2013-2014), the part that came back in counterparty for the work performed nationally reached a minimum of 31.5 percent of GDP, a very reduced value from the standpoint of the international practice. We ought to mention that employees’ remuneration is defined as the total compensation, in cash or in kind, payable to the employee by the employer in exchange for the work done by the former. It also includes the social contributions paid by the employer, in other words it considerably influences the level of pensions paid from these social contributions for the past work.

Table 16: Remuneration of Private Labour in 1995-2014

State Remuneration of Employees Wage (% Total Year GDP (mill. euro) (% of GDP) Share Remuneration) (% GDP) 1995 7,611.70 2,858.40 37.6 5.7 15.1 1996 11,463.50 4,318.00 37.7 4.9 12.9 1997 25,689.10 8,270.70 32.2 4.8 15 1998 37,257.90 14,087.10 37.8 5 13.3 1999 55,479.40 19,017.00 34.2 7.5 22 2000 81,275.30 31.873,7 39.2 7.8 19.8 2001 118,327.20 48.604,9 41.1 6.6 16 2002 152,630.00 60.172,9 39.4 6.8 17.4 2003 198,761.10 73.674,2 37.1 7.7 20.1 2004 248,747.60 91.039,3 36.6 7.5 20.5 2005 290,488.80 112,996.90 38.9 8.2 21 2006 347,004.30 132,194.90 38.1 8.6 22.6

150 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

State Remuneration of Employees Wage (% Total Year GDP (mill. euro) (% of GDP) Share Remuneration) (% GDP) 2007 418,257.90 153,941.00 36.8 8.5 23.2 2008 524,388.70 205,876.40 39.2 9.1 23.2 2009 510,522.80 192,983.20 37.8 9.9 26.2 2010 533,881.10 189,790.90 35.5 8.5 23.9 2011 565,097.20 185,462.60 35.5 7.1 21.6 2012 596,681.50 195,223.40 32.7 7.1 21.7 2013 637,583.10 200,848.60 31.5 7.2 22.7 2014 669,509.20 211,189.30 31.5 7.1 22.4

Source: NIS, personal calculations For reference, the average of employee wages in GDP at EU 28 level was equal to 48 percent in 2013. Among the states from our region, Poland recorded the lowest percentage, equal to 37 percent while Hungary, the Czech Republic and Bulgaria exceeded 40 percent. In the great economies of the EU, such as France, Great Britain and Germany, values of over 50 percent were recorded and Italy was the only one around 40 percent. Within the total remuneration, the share of those working in public administration and defence, social insurances in the public system, education, health and social assistance etc. reached the minimum of 12.9 percent in 1996. In terms of share in GDP, 1997 was the year in which the percentage related to state wage dropped to 4.8 percent.

It is important to notice that the minimum values for wages were reached during crisis years, 1997 (32.2% of GDP), 1999 (34.2% of GDP) and when the crisis was ending, 2012 (32.7% of GDP) and 2013-2014 (31.5% of GDP each). These values show which countries were burdened by the crisis and which did (not) benefit from coming out of the difficult economic situation. There is, however, a fundamental difference between 1997 and the other years previously mentioned. In the first case, state employees suffered together with the others, proportionally, and nothing was given to them if there were not any resources available. Instead, in all the other cases, the state wage share in the total remuneration was around 22 percent and the share in GDP exceeded seven percentages, so it was reasonable in relation to the economic situation and the budgetary receipts.

Central Bank Journal of Law and Finance, No. 1/2015 151 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Table 17: Remuneration Share (% in GDP)

Year Total State Private 1997 32.2 4.8 27.4 1999 34.2 7.5 26.7 2012 32.7 7.1 25.6 2013 31.5 7.2 24.3 2014 31.5 7.1 24.4

Source: NIS, personal calculations The problem was in the private sector, where the piece of the wellness cake had systematically and considerably got thinner until it reached the damage threshold of 25 percent, which explains, at least partially, the problems encountered in workforce, in paying debts to banks and, in general, in making people engage in general objectives such as reducing development disparities and adopting the euro. If we look back on the start of the uninterrupted period of economic growth, we can see that the throttle was hit exactly on this segment of the share of work remuneration from the private sector in GDP. As long as the remuneration level was kept at around and above 30 percent (even during state wages exaggerations, which were about to exceed 10 percentages of GDP), results were good.

Table 18: Remuneration of Work on Total and Sectors (% of GDP) in 2000-2011

Year Remuneration State Private 2000 39.20 7.80 31.4 2001 41.10 6.60 34.5 2002 39.40 6.80 32.6 2003 37.10 7.70 29.4 2004 36.60 7.50 29.1 2005 38.9 8.2 30.7 2006 38.1 8.6 29.5 2007 36.8 8.5 28.3 2008 39.2 9.1 30.1 2009 37.8 9.9 27.9 2010 35.5 8.5 27 2011 35.5 7.1 28.4

Source: NIS, personal calculations

152 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

We should remember that the maximum level of 34.5 percent was reached in 2001, after some measures that massively changed the fiscal policy. More interesting, that is the year when the maximum was reached for work and at the same time there was a drop to a reasonable value (related to GDP) for state work, which created space for wages similar to the European model of work in the private sector. Somehow paradoxically, the data recorded since 2012 show that the tendency towards remunerating the work done has mostly disappeared (quite abruptly, according to official statistics), especially in the private sector, the one constituting the great majority and giving the general trend in economy. All that it is left is to be puzzled about the low potential of growth and about the reason why Romania does not perform better.

8. ROMANIA – COMPLETELY OUTSIDE THE EUROPEAN UNION AS FAR AS THE NUMBER OF EMPLOYEES IS CONCERNED

There is an objective motivation of the more reduced work remuneration share in GDP. The official number of employees is equal to half of the lowest value recorded in the EU states, according to the last data provided by Eurostat. Almost all the European states work with a ratio of employees/citizens of 40%-50%, while in our country this ratio is somewhere below 30%.

Figure 1: Number of Employees/Citizens (%) in the European States Source: Eurostat

Central Bank Journal of Law and Finance, No. 1/2015 153 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

A clear distinction can be made between three major groups of countries, according to their position near the upper or lower threshold of the waged work: 1. “50% type” countries with an employee per two citizens: Netherlands, Denmark, Sweden, Austria, Great Britain, Germany, Bulgaria. 2. “40% type” countries with two employees per five citizens: Greece, Spain, Ireland, Italy, France, Romania was in this category in 1989, as well. 3. “30% type” Romania of the year 2015, far away from the other EU members, although Romania is supposed to have a functioning market economy. Certainly, given the concrete circumstances, that does not necessarily mean that in the first group people work more. However, the work relations are recorded, fact which enables their taxation according to the current legislation, thus assuring budgetary sources for pensions and social protection. If some former socialist countries such as Hungary, Slovakia, Estonia and even Poland are placed at “Latin levels” of employees’ share, others such as the Czech Republic or Bulgaria get close to the levels from the north of the continent. Note that the richest country in Europe, Norway, outside the EU, has the highest percentage of employees in the total population, namely 54 percent, level given for comparison in the same statistics of the EU. Only two countries are not present with data in statistics: Luxembourg and Romania. However, the reasons are completely different. The workforce in the small Luxembourg mostly belongs to the neighbouring countries, which complicates the statistics and explains, among others, the exceptional standard of living three times higher than the EU average (many people work, but the result is only divided by the number of residents). In Romania, the comparison methodology has not been well set up and the reference basis remains uncertain. The number of 5.2 million employees leads to a very low result from a European perspective. Even if we took into account the numbers presented by NIS – 6.097 thousand employees – which also include military services, assimilated persons, the informal sector and the persons working on the black market and if we considered a number of 19 million citizens (according to the most recent census and to the European norms of residence and travelling), we would still not reach the level of 30 percent. This level, beyond the practice and the EU, constitutes a major problem for our country.

154 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

Figure 2: Countries with a Waged Work Comparable to Romania Source: Eurostat

9. WHERE DID THE INVESTMENTS IN THE “NEW MAN” GO? STATISTICS OF PEOPLE WHO LEFT THE COUNTRY BASED ON AGES AND CONSEQUENCES

One in five Romanians born during the decade of the revolution from December has left the country, according to the data published by the National Institute of Statistics. The “new man” has come, in considerable percentages, to support other countries’ development and to contribute to their systems of social security, because we were not careful enough to ensure jobs for “the most precious capital”. The percentage of Romanians who went abroad more than 12 months ago (there are also numerous cases in which people commute seasonally and live in the country off the money gained abroad) has risen to almost 20 percent, if we calculate the values estimated by NIS for the age segment of 25-39 years old, for the resident population and the one that left the country.

Central Bank Journal of Law and Finance, No. 1/2015 155 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Table 19: Number of Romanians According to Residence

Age Group Total Residents Emigrants % Emigrants 0 – over 85 years old 22 286 825 19 942 642 2 344 183 10.5 15 – 24 years old 2 582 216 2 277 295 304 921 13.4 25 – 39 years old 5 745 400 4 687 659 1 057 741 18.4 40 – 49 years old 3 371 174 2 960 910 410 264 13.8 Source: NIS, personal calculations The majority of the people gone abroad is part of the generation born and raised with great sacrifices in the ‘80s, the toughest period during Ceaușescu’s ruling. Without any future in their country, without jobs and without the possibility to purchase a place to live, a great part of these people took advantage of the openness towards the West and headed to places where their skills were more appreciated and fairly rewarded, where crediting systems allow them to develop a business or purchase a place to live.

Table 20: Number of Romanian Residents and Emigrants

Age (Years) Total Residents Emigrants % Emigrants 25 358,668 288,467 70,201 19.6 26 362,138 289,341 72,797 20.1 27 354,982 282,994 71,988 20.3 28 340,460 267,968 72,492 21.3 29 329,152 262,438 66,714 20.2 30 313,499 241,789 71,710 22.8 31 334,678 257,984 76,694 22.9 32 361,977 284,450 77,527 21.4 33 375,964 299,422 76,542 20.3 34 382,369 306,185 76,184 19.9 Source: NIS, personal calculations The fact is that those who left the country were exactly the people who could have had a significant contribution to the creation of added value; they are the people who supported their retired parents (whose needs were difficult to cover from the taxes and dues of those who worked in Romania) and, indirectly, the national currency. Nonetheless, given the investments in raising and educating the people who left the country, which did not pay off in the long term, but materialized in temporary and discretionary currency entry, their strictly material result for the society they left seems to be negative, with multiple consequences.

156 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

If we look at the structure on age and gender, we also have an explanation regarding the considerable decrease in the number of births. The number of women aged 18-30 who left the country is higher than the number of men abroad.

Table 21: Demographic Structure per Age and Gender

Age Total Male Female Female–Male 18 19,719 9,784 9,935 +151 19 21,850 10,598 11,252 +654 20 26,493 12,796 13,697 +901 21 30,625 14,256 16,369 +2,113 22 38,710 17,835 20,875 +3,040 23 51,031 23,391 27,640 +4,249 24 61,334 28,098 33,236 +5,138 25 70,021 32,614 37,587 +4,973 26 72,797 33,349 39,448 +6,099 27 71,988 33,972 38,016 +4,044 28 72,492 34,711 37,781 +3,070 29 66,714 32,907 33,807 +900 30 71,710 35,865 35,845 -20 31 76,694 39,898 36,796 -3,102 32 77,527 40,443 37,084 -3,359 33 76,542 40,828 35,714 -5,114 34 76,184 41,205 34,979 -6,226 35 74,703 41,030 33,673 -7,357 Source: NIS, personal calculations Obviously, there are also other explanations for the services performed in destination countries (see the last column in the table). Clearly, we will become fewer and fewer, with a smaller pension. Basically, developed economies have mainly covered the directly productive jobs with their own citizens and have allowed recently officialised Europeans in less attractive or more difficult sectors for the local population. Note that for the officially declared unemployment level for February 2015 of 5.5 percent and approximately half a million persons registered as such in the country, we have around a million and a half Romanians abroad. Obviously, they have to make a living somehow and most of them work.

Central Bank Journal of Law and Finance, No. 1/2015 157 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

Table 22: Structure of Unemployment at the End of February 2015

Age Group Number of Persons Under 25 years old 64,834 25 – 39 years old 151,664 40 – 49 years old 142,550 50 years old and above 139,483 Total 498,531 Source: National House of Pensions and Other Social Insurance Rights Under these circumstances, if it had not been for the access to the European Union and the corollary of the free circulation of workforce (the transition period of seven years in which Member States imposed temporary restrictions for the workforce from Romania is over), we would have had a totally different level of unemployment, maybe not quite equal to 20 percent according to the rule of cross-multiplication, but at least triple. This reality starts from the incapacity of the local market economy (as it became during the transition period) to ensure enough jobs and from the insufficient incentives given by governors in order to create jobs, hence the impossibility to reduce faster the development disparities. Plain and simple, beyond the situation from investments and technology, but in correlation with them, the lack of workforce considerably reduces our potential GDP, namely the rhythm in which we could grow, without further relapses and without a re- inflammation of the price rise, over a longer period of time. The deficit in the human factor in terms of generations has affected our long-term balance of the social security system, namely both the actual level of payable pensions and the perspectives of providing the ones who stayed in the country with a reward correlated with the support they also offered during the difficult transition towards the market economy.

10. ACTUAL INDIVIDUAL CONSUMPTION AND EURO ADOPTION

In 2013, Romania reached 57 percent of the average standard of living of EU 28, in the version of the actual individual consumption (AIC), according to the data released by Eurostat. In the classical version of the standard of living monitoring, which measures GDP per capita at the standard purchasing power parity, Romania was at 55 percent of the Union average. The actual individual consumption (AIC) indicator reflects a citizen’s wellness in a certain country better than GDP per capita. This last indicator, more often used and known to the public, is set based on purchased goods and services paid by a household.

158 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean

Instead, AIC is calculated based on the actual goods and services consumed by individuals, without taking into account that they were contracted and paid by households, government or non-profit organizations. Therefore, several opinions show that it is preferable for international comparisons and that it reflects the standard of living more accurate and reliable. The differences occur when the percentage of payments made directly by households for important services, such as health and education, are very different from one country to another. Thus, the apparent order between countries of the standard of living that would result from the classical GDP per capita can bear important modifications. We presented the values from 2011-2013 to show the evolutions in the last three years analysed by Eurostat. Moreover, for the correct perception of the European context and the positioning in the region we are part of, we presented the values from several countries outside EU. Romania is part of the small group of states, which registered relative increases in the EU28 average at the level of GDP per capita and AIC per capita, along Lithuania, Poland, Slovakia, Estonia and Latvia. Note that the distance separating us from Bulgaria has increased considerably (eight percentages in AIC and ten percentages in GDP per capita).

Table 23: AIC per Capita and GDP per Capita in 2011-2013 in the EU Member States and other Countries in Proximity AIC per capita GDP per capita Country 2011 2012 2013 2011 2012 2013 EU28 100 100 100 100 100 100 Eurozone 107 106 106 108 108 107 (EA18) Luxembourg 138 139 136 265 264 257 Germany 121 122 122 122 123 122 Austria 120 120 120 128 129 128 Denmark 114 115 115 126 125 124 Finland 113 115 115 117 115 113 Sweden 114 114 115 127 126 127 Great Britain 114 115 115 106 107 109 Netherlands 118 116 113 135 132 131 Belgium 111 112 111 120 120 119 France 111 110 110 108 107 107 Italy 106 102 100 103 101 99 Ireland 97 94 94 130 130 130 Cyprus 98 98 93 96 93 89 Spain 91 90 90 95 94 94 Portugal 83 80 84 78 76 79 Greece 89 84 83 77 74 73 Malta 80 80 79 84 84 86

Central Bank Journal of Law and Finance, No. 1/2015 159 The Evolution of GDP in Romania and the Consequences from the Purchasing Power Perspective

AIC per capita GDP per capita Country 2011 2012 2013 2011 2012 2013 Lithuania 70 73 78 65 69 73 Slovenia 79 78 76 80 81 79 Poland 70 73 75 64 66 67 Czech 73 73 74 83 82 82 Republic Slovakia 72 73 74 73 74 75 Estonia 59 63 65 68 71 73 Latvia 57 60 65 57 60 64 Hungary 62 62 62 65 65 66 Croatia 59 60 61 60 61 61 Romania 53 55 57 51 53 55 Bulgaria 47 50 49 44 45 45 Norway 132 134 135 182 190 186 Switzerland 126 129 130 159 162 163 Iceland 113 115 116 115 116 119 Turkey 58 57 59 51 52 53 Montenegro 52 51 51 41 39 40 Serbia 45 46 46 36 37 37 Macedonia 38 39 40 34 34 36 Albania 32 33 33 28 28 28 Bosnia- 36 37 37 28 28 29 Herzegovina Source: Eurostat The most spectacular difference between AIC and GDP per capita indicators at EU level is noticed in the case of Ireland, which is in top according to GDP per capita, but under the European average in terms of the actual individual consumption (Luxembourg is a special case due to the afflux of workers from the neighbouring countries). Italy, the second great commercial partner of Romania, best represents the EU average in terms of reference and standard of living indicators for our process of real convergence. However, Italy is present at EU28 level in an isolate manner, quite far from the bloc of developed states, or placed above and far from the bloc of states dealing with economic problems, which are below Italy as far as the standard of living is concerned. The first bloc comes after the prize- winners, German speaking countries, and has its limits set by Northern countries plus Great Britain and France, among which there are also Netherlands and Belgium. The second includes Ireland, along with the southern Cyprus, Spain, Portugal and Greece. Coming back to the ratio between AIC and GDP per capita, note the superior positioning of Great Britain compared to Netherlands at the level of goods and services actually consumed (115 percent to 113 percent of EU average), although the difference at the level of GDP per capita is considerable (109 percent compared to 131 percent of EU average). Closer to us, Poland overtakes the Czech Republic in a similar situation (75 percent to 74

160 Central Bank Journal of Law and Finance, No. 1/2015 Dan Pălăngean percent of EU average in terms of actual consumption compared to a definitely unfavourable “score” in terms of GDP per capita, respectively 67 percent to 82 percent). The distance separating Romania from Hungary at the level of GDP per capita is cut in half if we take into account the AIC indicator. Finally, as far as the convergence necessary for the euro adoption is concerned (the limit of 60 percent of the GDP per capita EU average), if we consider a more adequate AIC, there are two important observations. The first one: the most recent member of the Eurozone, Lithuania adopted euro at an AIC of approximately 80 percent (it had already been 78 percent a year before). The second: in the year of accession (2011), Estonia had an AIC of 59 percent, only 2 percent higher than our country in 2013. Latvia, whose two indicators are more balanced, accessed the Eurozone in 2014 with an AIC equal to 65 percent and a GDP per capita of 64 percent, significantly above the value of 60 percent we were supposedly setting as a target. In other words, one can find enough arguments pro and against the euro adoption from 2019, which turns it into a completely political decision (if we relate to the dry data for Latvia and Lithuania), focused on the support of the population (based on Estonia’s experience).

11. CONCLUSIONS

The analysed data clearly show the costs of transitioning between economic systems and confirm the famous prediction of the late political analyst Silviu Brucan that it would take the Romanian society 20 years to undergo a real change. Practically, a sacrificed generation absorbed the costs of re-establishing the economy on its capitalist bases, while we still had the chance of integration into the EU under favourable circumstances. Although the overall macroeconomic data look good on paper, the public perception of western-type, civilized living, is established successively in Romanian regions, with a pronounced gap between the large urban centres and the rest of the country. The nationwide under-utilisation of workforce leads to the phenomenon of migration towards other European states, with gains on a personal level, but with a negative impact on the country’s development pace, also taking into account the damaging effect on the social security system sustainability. Only from now on, we will be able to maximize our chances to a European standard of living and to make the most out of the free movement of the factors of production within the Union. The next step is joining the unified European currency family, which was pledged at the moment of adherence and is already announced at a political level. This will be the maturity test for both Romanian economy and society, towards a real convergence, subject to the inherent rigors of the complex process of harmonising with the EU peer countries.

Central Bank Journal of Law and Finance, No. 1/2015 161