Closure within multi-plant firms: evidence from Japan

Abstract

In the literature there is substantial evidence that a plant is more likely to be closed down if it is owned by a firm with other plants or is owned by a multinational enterprise (MNE). But does ownership or multi-plant status matter for which plants are closed? Using Japanese data we study plant closure by multi-plant MNEs and non-MNEs. We show that both organisational forms raise the probability of plant exit and that plants that are relatively small and capital unintensive relative to the rest of the firm are significantly more likely to exit.

Appendix

1.1 Total factor productivity

There are 48 manufacturing industries in our data set. Total factor productivity (TFP) is calculated for each plant relative to the industry average. Following Aw et al. (1997), we define the TFP level of plant i in year t in a certain industry, j, in comparison with the TFP level of a hypothetical representative plant in year 0 in that industry as follows

$$ \begin{aligned} \ln TFP_{ijt} &= \left( {\ln Q_{it} - \overline{{\ln Q_{t} }} } \right) - \sum\limits_{i = 1}^{n} {\frac{1}{2}\left( {S_{zit} + \overline{{S_{zt} }} } \right)} \left( {\ln X_{zit} - \overline{{\ln X_{zt} }} } \right) \hfill \\ & \quad + \sum\limits_{s = 1}^{S} {\left( {\overline{{\ln Q_{s} }} - \overline{{\ln Q_{s - 1} }} } \right)} - \sum\limits_{s = 1}^{S} {\sum\limits_{z = 1}^{Z} {\left( {\overline{{S_{zs} }} + \overline{{S_{zs - 1} }} } \right)\left( {\overline{{\ln X_{zs} }} - \overline{{\ln X_{zs - 1} }} } \right)} } \hfill \\ \end{aligned} $$

(4)

where Q _it , S _zit and X _zit denote the gross output of plant i in year t, the cost share of factor z for plant i’s input of factor z in year t. Variables with an upper bar denote the industry average of that variable. We use 1994 as the base year. Capital, labour and real intermediate inputs are used as factor inputs.¹⁸

The representative establishment for each industry is defined as a hypothetical plant whose gross output as well as input and cost share of all production factors are identical to the industry average. The first two terms on the right-hand side of Eq. (4) denote the gap between plant i’s TFP level in year t and the representative plant’s TFP level in year t and the representative plant’s TFP level in the base year. lnTFP _ijt in Eq. (4) constitutes the gap between plant i’s TFP level in year t and the representative plant’s TFP level in the base year.

1.2 Industry variables

Globalisation has been shown to cause exit. The source of import competition in the US affects plant survival and causes firms to adjust their product mix (Bernard and Jensen 2002; Bernard et al. 2006). We disaggregate import penetration into low-wage import penetration (LWPEN) and import penetration from all other countries (OTHPEN).¹⁹ These measures are calculated as:

$$ LWPEN_{jt} = \frac{{M_{jt}^{LW} }}{{M_{jt} + Y_{jt} - X_{jt} }};\,OTHPEN_{jt} = \frac{{M_{jt} - M_{jt}^{LW} }}{{M_{jt} + Y_{jt} - X_{jt} }} $$

where LWPEN _it represents low-wage country import competition in industry j at time t, M ^LW _jt is the value of imports from low-wage countries in industry j at time t, M _jt and X _jt represents the value of total imports and exports in industry j at time t and Y _jt denotes output in industry j during year t. OTHPEN _jt denotes imports from all countries except low-wage economies.

Intra-industry trade is often found to have a positive effect upon firm exit. As international trade grows firms diversify their product range which may lead them to enter new industries and exit sectors they operate in currently. Our measure of intra-industry trade is constructed using the Grubel–Lloyd index:

$$ GL_{jt} = \left[ {\left( {X_{jt} + M_{j} } \right) - \left| {X_{jt} - \left. {M_{jt} } \right|} \right.} \right]\frac{100}{{\left( {X_{jt} + M_{j} } \right)}} $$

where GL _jt is the Grubel–Lloyd index of intra-industry trade in industry j in year t, X _j are exports in industry j during year t and M _jt are imports in industry j during year t (Table 6).

Table 6 Industry variables

The industry variables mentioned so far capture the influence of globalisation upon plant exit. We also include a measure of sunk costs. The empirical literature has identified sunk costs as being an important factor in shaping exit. Sunk costs also play a key role in determining exporting behaviour (Roberts and Tybout 1997) and can affect the distribution of productivity in the industry (Aw et al. 2003). We define sunk costs as the minimum of the entry and exit rate in the plant’s 3-digit industry, that is²⁰

$$ sunk \, costs_{jt} = - \left\{ {\min \left[ {entry_{jt} ,exit_{jt} } \right]} \right\}.$$