Export characteristics and output volatility: comparative firm-level evidence for CEE countries

Abstract

The literature shows that openness to trade improves long-term growth but also that it may increase exposure to high output volatility. In this vein, our paper investigates whether exporting and export diversification at the firm level have an effect on the output volatility of firms. We use large representative firm-level databases from Estonia, Hungary, Romania, Slovakia and Slovenia over the last boom–bust cycle in 2004–2012. The results confirm that exporting is related to higher volatility at the firm level. There is also evidence that this effect increased during the Great Recession due to the large negative shocks in export markets. We find that export diversification mitigates volatility only in some cases. Exporting more products or serving more markets does not necessarily result in higher stability of firm sales.

Appendices

Appendix 1: Theoretical model of export diversification and output volatility

Firm i produces output Y _it using the Cobb–Douglas production function with constant returns to scale and with domestic labour L _it with share $\alpha$ and domestic capital K _it as inputs. Producers’ technology, A _it is subject to firm-specifics shocks. Firm sells a time-invariant share, 1 − λ _i of output domestically for price p _t and exports the rest. The firm can sell to one or more foreign markets m for $p_{mt}^{*}$ price, with the sum of foreign market shares, $\mathop \sum \nolimits_{m = 1}^{M} \lambda_{mi}^{*}$, totalling λ _i . Exporting cost per unit of product for each destination market, c _mt . Firms’ maximisation in the equilibrium yields:

$$\ln \bar{Y} = \beta_{0} + \beta_{1} \ln \bar{A} + \beta_{2} \ln \bar{d} + \beta_{3} \ln \bar{K}$$

(2)

Where denoting the labour supply and demand elasticities with $\eta^{\text{S}}$ and $\eta^{\text{D}} :\;d_{it} = \left[ {p_{t} (1 - \lambda_{i} ) + \mathop \sum \nolimits_{m = 1}^{M} \left( {p_{mt}^{*} - c_{m} } \right)\lambda_{mi}^{*} } \right]$, $\beta_{0} = \alpha \left[ {\eta^{D} \eta^{S} /\left( {\eta^{D} + \eta^{S} } \right)} \right]ln\alpha$, $\beta_{1} = 1 + \alpha \left[ {\eta^{D} \eta^{S} /\left( {\eta^{D} + \eta^{S} } \right)} \right]$, $\beta_{2} = \alpha \left[ {\eta^{D} \eta^{S} /\left( {\eta^{D} + \eta^{S} } \right)} \right]$, $\beta_{3} = 1 - \alpha$. Variables with bars denote equilibrium values and time subscripts suppressed.

We assume the firm faces random technology, demand and capital cost shocks, $\gamma_{A} ,\;\gamma_{d} \; {\text{and}}\;\gamma_{K}$ respectively. The variables can be then defined as $= \bar{A}e^{{\gamma_{A} }} ,d = \bar{d}e^{{\gamma_{d} }} = \bar{d}e^{{\gamma_{p} \left( {1 - \lambda } \right) + \mathop \sum \limits_{m} \gamma_{{p_{m}^{*} }} \lambda_{m}^{*} }}$, and $K = \bar{K}e^{{\gamma_{K} }}$, where the demand shock depends on the domestic and foreign demand shocks.¹⁷ Deriving the expressions of output with random shocks over the equilibrium output, we obtain the following expression for the variance of the output growth.

$$\begin{aligned} Var\left( {\ln \left( {\frac{{\hat{Y}}}{Y}} \right)} \right) = & \beta_{1}^{2} Var\left( {\gamma_{A} } \right) + \beta_{2}^{2} \left[ {\left( {1 - \lambda } \right)^{2} Var\left( {\gamma_{p} } \right) + \mathop \sum \limits_{m} \lambda_{m}^{*} {}^{2}Var\left( {\gamma_{{p_{m}^{*} }} } \right)} \right] + \beta_{3}^{2} Var\left( {\gamma_{K} } \right) \\ & \quad + \,2 \times \beta_{2}^{2} \left[ {\mathop \sum \limits_{m} \left( {1 - \lambda } \right)\lambda_{m}^{*} Cov\left( {\gamma_{p} ,\gamma_{{p_{m}^{*} }} } \right) + \mathop \sum \limits_{m \ne j} \lambda_{m}^{*} \lambda_{j}^{*} Cov\left( {\gamma_{{p_{m}^{*} }} ,\gamma_{{p_{j}^{*} }} } \right)} \right] \\ \end{aligned}$$

(3)

If a firm is producing for one market only, be it domestic or foreign, the variance of output depends on three components: variance of productivity shocks, variance of demand shocks in the market, and variance of shocks to capital. If a firm is producing not for one but for M markets, then the output volatility equation is composed of $\left( {{\text{M}}^{2} + 2} \right)$ components, which are two variances representing the technology and capital cost shocks, M variances to the demand shocks, and $\frac{{{\text{M}}\left( {{\text{M}} - 1} \right)}}{2}$ double covariances between the demand shocks for all the markets. Due to the complexity of relations, entry to an additional market could either decrease output volatility because of the diversification effect in the covariance terms, or increase the output volatility because of the high variance or covariance of demand shocks of a firm in additional markets.

Appendix 2: Diversification of products exported and destination markets in 2008 and 2012

See Fig. 3.

Fig. 3

Source: Authors’ calculations from CompNet and customs data

Diversification of products exported (left panel) and destination markets (right panel), manufacturing, 2008 and 2012. Note: The figures are inspired by the presentation by Arkolakis and Muendler (2013) of exporter scope distribution.

Appendix 3: Estimates of concentration and volatility over different time spans, 2004–2012

See Table 11.

Table 11 Export concentration and output volatility with additional controls, rolling window between 2004 and 2012.

Appendix 4: Estimates of concentration and volatility over different time spans, alternative measure of concentration, 2004–2012

See Table 12.

Table 12 Export concentration and output volatility with additional controls, alternative measure of concentration, rolling window between 2004 and 2012.

Appendix 5: Estimates of concentration and volatility over different time spans, volatility over 6 years, 2004–2012

See Table 13.

Table 13 Export concentration and output volatility with additional controls, volatility over 6 years, rolling window between 2004 and 2012.