Business Cycle Accounting: Bulgaria after the introduction of the currency board arrangement (1999-2014).

Author:Vasilev, Aleksandar
  1. Introduction and Motivation

    This paper focuses on explaining the economic fluctuations in Bulgaria after the introduction of the currency board arrangement in 1997, the period of macroeconomic stability that ensued, the EU accession, and the episode of the recent global financial crisis. In this way the current study tries to fill an important gap in the literature on transition economies: More specifically, one objective of the study is to show that Bulgaria's transition and EU accession experience, as well as the slump during the financial crisis are comparable in magnitude to what other old and new EU member states experienced. (1) Therefore, it will be shown that modern business cycle accounting methodology, as in Chari et al. (2002, 2007), when applied to Bulgaria can produce results which are comparable with findings in other countries. (2)

    Over the period studied in this paper (1999-2014), Bulgarian governments implemented a lot of structural reforms that achieved macroeconomic stability, fulfilled all the accession criteria and the country joined the EU as a new member state. (3) First, following an episode of a banking and financial crisis, in mid-1997 Bulgaria fixed its exchange rate by setting its currency, Bulgarian lev (BGN), initially at par to the German mark, and from 2001 onwards--to the Euro (1 Euro = 1.95583 BGN) and liberalized most of its markets, and continued privatizing state assets. Government finances were put in order, subsidies to loss-making state enterprizes, as well as any form of state aid, were discontinued, and the economy started re-orienting its export markets to Western European countries, increasing trade with the other EU member states at the expense of former markets in the East, such as Russia. The EU structural and accession (or, "cohesion," as they are also called) funds also helped Bulgaria develop its infrastructure, which decreased transportation costs, attracted foreign investors and increased inter-city labor mobility. The EU accession process and the change in trade partners are easy to be incorporated within the neoclassical paradigm--trade shocks would be measured as productivity shocks in a closed-economy setup when one regards trade as a technology that allows countries to convert inputs into outputs at a lower opportunity cost.

    As seen from Figure 1 on the next page, the period covered in this study (19992014) exhibits a typical business cycle pattern: until 2004 Bulgarian output per capita is below trend, then between 2004-08 the economy is above trend. (4) As seen from the plot, the global financial crisis hits Bulgaria in 2009, and as of 2014, the economy still has not fully recovered from the slump.

    Next, before we proceed with the business cycle accounting, we start by decomposing the growth rate in Bulgaria into its elements in order to understand the major forces at work during the last fifteen years, and as a motivation for the computational experiment to be performed. The results from the growth accounting procedure are presented in the next section. The rest of the paper is organized as follows: Section 3 describes the model and the equilibrium concept to be utilized in the paper. Section 4 discusses the data to be used and describes the model calibration procedure. Section 5 solves for the steady-state. Section 6 computes the wedges using both the model and data. Section 7 conducts the counterfactual analysis. Section 8 provides some policy recommendations, and Section 9 provided conclusions.

  2. Growth Accounting

    Next, as an additional motivation of our study, we follow Prescott (2002) and Cavalcanti (2007): We define the aggregate production function as a Cobb-Douglas one with a time-varying trend, i.e. [Y.sub.t] = ([A.sub.t][[gamma].sup.t]) [K.sup.[alpha],t] ([N.sup.h.sub.t t]).sup.1-[alpha]], where [Y.sub.t] is aggregate output, [A.sub.t] denotes the level of Total Factor Productivity (TFP), [gamma] is one plus the average growth rate, [K.sub.t] is aggregate physical capital stock, [N.sub.t] is population, and [h.sub.t] are per person hours. We can take natural logs from both sides and thus decompose the log output into the following factors:

    In [y.sub.t] = t ln [gamma] + ln [A.sub.t] + [alpha]/1 - [alpha]] ln [k.sub.t]/[y.sub.t] + ln [h.sub.t], (1)

    where the small-case letters denote per capita variables. The first factor driving output per capita growth is the trend, the second is the technology (productivity) factor, the third is the capital-to-output ratio (the capital factor), and the forth one is the labor factor. (5) Those are plotted in Figure 2 below, where the index value for all variables is normalized to 100.

    As seen from Fig. 2, the weighted capital-to-output ratio is not varying much, thus leaving productivity changes and increases in the labor input as the major drivers of growth in output. (6) The same pattern is observed in Table 1 on the next page, yet another argument that the investment wedge is not going to be an important explanatory factor of business cycle fluctuations in Bulgaria over the period covered in this study.

    During the period 2000-2007 there is a lot of disinvestment, and after the EU accession mechanizing production happens at the expense of laying off redundant labor. The effect of the financial crisis that hit Bulgaria in 2009, and the output plunge is seen in the drop in productivity. From 2012 onward, there has been some recovery, but without growth in the labor input.

    The results from the growth accounting exercise in Table 1 on the next page suggest that, as in Cavalcanti (2007) for the case of Portugal, Bulgarian economy could also benefit greatly from labor market reforms as a way to improve output per worker. Policy measures could include (i) increasing in labor efficiency, (ii) decreasing the distortions resulting from labor market policies already in place, and (iii) introduce more flexibility in labor contracting, especially when it comes to the process of collective bargaining over wages and employment. The other set of reforms should be aimed at increasing overall productivity, which is broadly connected to the level of competition in major industries, the level of barriers to entry and exit, the rate of innovations and technology adoption, and institutional quality (e.g. as suggested in Prescott 2002). (7)

    Next, we will implement business cycle accounting for Bulgaria, as pioneered in Chari et al. (2002, 2007). The method consists of two major components: The first component is an equivalence result, which states that a large class of dynamic general equilibrium setups are equivalent to the stochastic optimal growth model with a representative agent and time-varying "wedges," which could be regarded as representing certain distortions and frictions that have an impact on overall economic efficiency and the allocations of the two major factors of production, capital stock and labor hours. Besides its tractability, the usefulness of the simple representative agent model is that it can be regarded as being an isomorphic representation to a much larger class of economic models, including but not limited to, much more sophisticated models with heterogeneous entrepreneurs engaging in investment project while facing credit constraints, as well as other setups with asymmetric information in financial markets and other credit frictions. All those frameworks are shown by Chari et al. (2002, 2007) to be equivalent to the baseline optimal growth model with "wedges", or distortions, that are allowed to vary over time. (8) Those wedges enter the production function, the inter-temporal ("the consumption-Euler"), and the intra-temporal optimality conditions of the household ("consumption-vs-labor"), and look like fluctuating productivity processes, and time-varying taxes on labor and capital. In light of this model correspondence, the wedges are denoted as the efficiency, labor, and capital/investment wedge. Given the theoretical stochastic optimal growth model, we can derive the optimality conditions describing the choices that are made by the rational agents in the framework, and then use empirical data on those equations to construct all the wedges.

    Furthermore, the methodology pioneered by Chari et al. (2002) and then improved in Chari et al. (2007), is also very much in the spirit of Lucas (1980), who argues that theoretical models are to be interpreted as laboratories within which controlled computational experiments can be performed. More specifically, once we have estimated the series for the wedges in the stochastic optimal growth model, we can assess the quantitative impact of each individual wedge on output per person in Bulgaria during the period 1999-2014, as well as the combined effect of different wedges. This is the second important component of Chari et al. (2002)'s methodology, the accounting part. The procedure is conducted through a counterfactual experiment: the realized values of the wedges will be inserted in the model one at a time, and the other wedges would be held fixed at their steady-state level. In other words, the model economy would be subjected to the observed (sub)set of shocks. As pointed out in Kehoe and Prescott (2007), total factor productivity is taken as "external to the micro decision makers but not as invariant to policy" (p.3). (9) This ceteris paribus logic would allow us to quantitatively evaluate, or to account for, the relative importance of each wedge, as well as the role of a certain combination of wedges. In this way the general-equilibrium model is used as a diagnostic tool to guide researchers to the factors that need to be studied in more detail. For example, changes in institutional quality can affect not only the productivity wedge, but also labor and capital accumulation, and thus the labor and capital wedges as well.

  3. Model Setup

    This is a relative standard representative-agent model. The stand-in individual household works and...

To continue reading