R&D spillovers and employment: evidence from European patent data

Abstract

In this paper we investigate the role of patents in the relationship between R&D activity, spillovers and employment at the firm level. A reduced-form labour demand equation is estimated. Our analysis is based upon a dataset consisting of 879 R&D-intensive manufacturing firms worldwide for which information was collected for the period 2002–2010. We use data from all EU R&D investment scoreboard editions issued every year until 2011 by the JRC-IPTS (scoreboards). Since the innovation output of the industrial strategy of every firm is the number of patents, the main contribution to the existing literature is to investigate also the impact of patents/R&D ratio and patents/spillovers ratio on employment level. The empirical results suggest a significant impact of R&D spillover effects on company employment although the results differ substantially according to the spillover stock, which may considerably affect policy implications.

Appendices

Appendix 1: A theoretical model

In this section, in order to illustrate the background and interactions on which our empirical model is built, we follow Garcia et al. (2004), in which an entrepreneur who competes in a differentiated product market, minimises costs with a constant returns to scale in traditional input technology. The entrepreneur invests in R&D for process and product innovation and to obtain new patents. The innovations and patents, once made, enter the production function at the beginning of the next period, during which the entrepreneur adjusts the product price and employment, taking into account the new technology and the expected demand. Furthermore, we assume that, not only as in Garcia et al. (2004) and in Aldieri et al. (2014), the innovation effects on both the technology and demand function are embodied by the impact of the accumulated knowledge capital denoted by K, but also that patents, denoted by B, occur in a partnership of accumulated knowledge capital and employment. Let ${\text{c be the marginal cost and w}}$ the vector of input prices. Then we may state that c = c(w, K). Furthermore, if $p$ is the output price, Y the output, $L$ the employment level, μ the entrepreneur’s mark-up on the marginal cost, d ^e an index of market dynamics, and finally with $K_{R} , p_{R} ,B_{R} ,L_{R}$ respectively the rival firms’ accumulated knowledge capital, output prices, number of patents and employment level, we can state the following:

$$p = \left( {1 + \mu } \right)c\left( {w,K} \right)$$

(2)

$$Y = D\left( {d^{e} ,p,p_{R} ,K,K_{R} ,B,B_{R} } \right)$$

(3)

$$K_{R} = g\left( K \right)$$

(4)

$$B = B\left( {K,L} \right)$$

(5)

$$B_{R} = B_{R} \left( {K_{R} ,L_{R} } \right)$$

(6)

$$L = c_{L} \left( {w,K} \right)Y$$

(7)

$$p_{R} = \left( {1 + \mu_{R} } \right)c_{R} \left( {w_{R} ,K_{R} } \right)$$

(8)

In the above, conditionc _L represents the derivative of the marginal cost with respect to labour input (Shephard’s lemma), while $c_{R} , w_{R} ,\mu_{R}$ indicate respectively the marginal cost, vector input prices and finally the mark-up for rival firms. After some simple substitutions we may write:

$$L^{D} = c_{L} \left( {w,K} \right)D\left[ {d^{e} ,\left( {1 + \mu } \right)c\left( {w,K} \right),\left( {1 + \mu_{R} } \right)c_{R} \left( {w_{R} ,g\left( K \right)} \right),K,g\left( K \right),B\left( {K,L} \right),B_{R} \left( {K_{R} ,L_{R} } \right)} \right]$$

(9)

Hence, in order to determine the short-run impact of innovation on the employment level, after simple algebraic manipulations we may have:

$$\frac{{\partial L^{D} }}{\partial K} = \frac{{\partial c_{L} }}{\partial K}Y + c_{L} \left\{ {\frac{\partial Y}{\partial K} + \frac{\partial Y}{\partial p}\frac{\partial p}{\partial K} + \frac{\partial Y}{{\partial p_{R} }}\frac{{\partial p_{R} }}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K} + \frac{\partial Y}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K} + \frac{\partial Y}{\partial B}\frac{\partial B}{\partial K} + \frac{\partial Y}{{\partial B_{R} }}\frac{{\partial B_{R} }}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K}} \right\}$$

(10)

$$\frac{{\partial L^{D} }}{\partial L} = c_{L} \frac{\partial Y}{\partial B}\frac{\partial B}{\partial L}$$

(11)

The first term on the right-hand side of this expression [Eq. (10)] is the displacement effect, the second is the sum of more compensation effects: the first is the effect of output innovation on demand; the second is the effect on demand via the reduction of the cost through price reduction; the third is effect on demand of the change in the price of rival firms via the effect on innovation of rivals; the fourth captures the effect of the innovations of its rivals on demand, the fifth is the effect on demand via the increase in patents. Finally, the concluding term measures the effect on demand due to increases in patents of rivals as a result of the consequence that changes in innovations have on innovations of rival firms.

Taking into account Eqs. (10) and (11), by differentiating Eq. (9), we will have:

$$\frac{dL}{dK} = \frac{{\frac{{\partial c_{L} }}{\partial K}Y + c_{L} \left\{ {\frac{\partial Y}{\partial K} + \frac{\partial Y}{\partial p}\frac{\partial p}{\partial K} + \frac{\partial Y}{{\partial p_{R} }}\frac{{\partial p_{R} }}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K} + \frac{\partial Y}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K} + \frac{\partial Y}{\partial B}\frac{\partial B}{\partial K} + \frac{\partial Y}{{\partial B_{R} }}\frac{{\partial B_{R} }}{{\partial K_{R} }}\frac{{\partial K_{R} }}{\partial K}} \right\}}}{{1 - c_{L} \frac{\partial Y}{\partial B}\frac{\partial B}{\partial L}}}$$

(12)

Once more, as in Aldieri et al. (2014), assuming that, at the start of achieving innovation, the entrepreneur bargains wages w with unions, and taking price dynamics (changes in $\mu$ and μ _R ) into account, according to the new competitive environment due to innovation, using $z$ and z _R to denote other possible causes of changes on wages and mark-ups, we can add the following equations:

$$w = w\left( {z,K} \right)$$

(13)

$$w_{R} = w_{R} \left( {z_{R} ,K_{R} } \right)$$

(14)

$$\mu = \mu \left( {z,K} \right)$$

(15)

$$\mu_{R} = \mu_{R} \left( {z_{R} ,K_{R} } \right)$$

(16)

Hence Eq. (9) will become:

$$\begin{aligned} L^{D} = \hfill \\ c_{L} \left( {w\left( {z,K} \right),K} \right)D\left[ {d^{e} ,\left( {1 + \mu \left( {z,K} \right)} \right)c\left( {w\left( {z,K} \right),K} \right),\left( {1 + \mu_{R} \left( {z_{R} ,g\left( K \right)} \right)} \right)c_{R} \left( {w_{R} \left( {z_{R} ,g\left( K \right)} \right),g\left( K \right)} \right),K,g\left( K \right),B\left( {K,L} \right),B_{R} \left( {K_{R} ,L_{R} } \right)} \right] \hfill \\ \end{aligned}$$

(17)

from inspection of which the short-run impact of innovation on the employment level will be converted into the following condition :

(18)

.From inspection of Eq. (18) it may be detected that the introduction of the four conditions (Eqs 12–16) modifies both the displacement and the compensation effects.

Appendix 2

See Table 3.

Table 3 Sectorial and geographical characteristics of firms

Appendix 3

See Table 4.

Table 4 Descriptive statistics of variables by geographical area