Productivity, outsourcing and exit: the case of Australian manufacturing

Abstract

This paper uses a panel of small and medium manufacturing firms in Australia and studies the relationship between productivity and outsourcing accounting for the possibility of inefficient firms self-selecting into exit instead of outsourcing to domestic suppliers. Estimating a propensity model on an unbalanced panel of firms and correcting for the selection bias when firms opt for exit shows that the impact of productivity on the outsourcing decision could be much larger than estimated so far. The paper further explores the impact of outsourcing on a firm’s future performance and finds that the effect is non-uniform and productivity dependent, and outsourcing mostly brings improvements to firms that initially had low productivities. The most productive firms seem to outsource for other reasons such as focusing on innovation and exports with an eye on longer-term returns.

Appendices

Appendix 1: Estimates of production functions

The coefficients $\beta_l$ and $\beta_k$ in the production function (1) are estimated in two different ways: Levinsohn–Petrin and system GMM. In Table 13, I report the estimated coefficients for each subsector and using each methodology that are used to form the measures of productivity.

Table 13 The estimated coefficients in the production function by subsector

Appendix 2: Description of variables

Employment: :

The way employment is reported in the BLS changes over years. In 1994–1995 and 1995–1996, employment is reported as the total number of full-time (TOTFT) and part-time (TOTPT) non-managerial employees. TOTWPP reports the number of working owners and proprietors, and TOTMAN reports the total number of managerial employees. The number of managers in a firm is constructed as MAN=TOTMAN + TOTWPP.

In 1996–1997, the BLS further breaks the non-managerial staff by casual (a form of temporary informal employment in Australia) and others. Total number of full-time non-managerial employees is thus constructed as OEFT+CASFT, and the number of part-time non-managerial employees is OEPT+CASPT

Value of output: :

In the BLS, the nominal sales of goods and services are provided in variables SALES5–8. The BLS also reports the change in the stocks of finished goods by indicating the opening stock OPSTOCK5–8 and the closing stock CLSTOCK5–8. The value of output produced over the year is computed as

$$\begin{aligned} \text {OUTPUT} = \text {SALES}+\text {CLSTOCK}-\text {OPSTOCK}. \end{aligned}$$

Value added output: :

Is computed as

$$\begin{aligned} \text {VA} = \text {OUTPUT}-\text {PURCHAS}-\text {OTHEXP}, \end{aligned}$$

in which, PURCHAS is the purchase amount of material inputs reported in the BLS, and OTHEXP is the other operation expenses (including contracted out jobs).

Value of capital: :

In the BLS, the nominal stock of plant and machinery capital is reported in NCASSPL5–8. The value of leasing stock is reported in RLHX5–8. The total stock of capital is computed as

$$\begin{aligned} \text {CAPITAL} = \text {NCASSPL}+\text {RLHX}/(r+\delta ). \end{aligned}$$

Cost of material: :

Is the value of material purchased PURCHAS5–8 in the BLS.

OUTS: :

Dummy indicating if a firm contracted out jobs that used to be done by its own employees. Using BLS variable CONOUT5–7. Survey Question: During the financial year, did this business contract out any activities that were previously performed by this business’s employees? (Exclude contracting out solely to handle peaks in workload.)

EXPINT: :

Export intensity is formed by dividing nominal value of exports by the nominal value of sales. Using the BLS variables SALES5–8 and EXPORTS5–8.

U25–50: :

Dummy indicating if the proportion of union members in a firm is 25–50 %. The BLS reports the proportion in bins of 0–10 %, 11–25 %, 26–50 %, 51–75 % and 76–100 %. Using the BLS variable UNIONME5–8. Survey Question: Please estimate the percentage of persons working for this business who were union members as at 30 June. (Please tick one)

$\hbox {U}>50:$ :

Dummy indicating if the proportion of union members in a firm’s employment is more than 50 %. Using the BLS variable UNIONME5–8. Survey Question: See above.

STARTX: :

Firm’s intention to commence exporting. In 1994–1995 and 1995–1996, the BLS reports if the firm intends to commence or maintain exporting in the next three years. STARTX in those years is set to one if the firm has the intention and did not report any exports. In later years, the BLS directly reports if the firm intends to commence exporting. Using the BLS variables INTCOEX5–8 and EXPORTS5–6. Survey Question: Please indicate if this business intends to Commence exporting during the next 3 years.

INNOVAT: :

Dummy indicating if a firm introduced a substantially new product or production process during the year. Using the BLS variable INNOVAT5–8. Survey Question: For the financial year reported, did this business develop any new products, or introduce any substantially changed products, or develop or introduce any new or substantially changed processes?

MULTI: :

Dummy indicating if a firm has multiple locations. Using the BLS variable BUSLOCS5–8 that reports the number of business locations. Survey Question: Number of locations operated by this business as at 30 June.

AGE: :

The BLS reports the age of a firm in a few bins. The bins are 0–1, 2–4, 5–9, 10–19 and 20+. Dummies are set the same as these bins. Using the BLS variable AGE5–8.

DECISION: :

Is the same as the BLS variable MANDIREC and indicates whether a firm has a major decision maker. The question is asked only once and applied to the whole panel.

FAMILY: :

Is the same as the BLS variable FBWPP and indicates whether a firm is run by a family with family members working as owners or proprietors. The question is asked only once and applied tot he whole panel.

INCORP: :

Is the same as the BLS variable TOLO and indicates the type of legal organization: incorporated or unincorporated. The question is asked only once and applied tot he whole panel.