Unbundling Technology Adoption and tfp at the Firm Level: Do Intangibles Matter?

• Author: Filippo Belloc,Massimo Del Gatto,Michele Battisti
• Date: 01 June 2015
• Published date: 01 June 2015
• DOI: http://doi.org/10.1111/jems.12094

Unbundling Technology Adoption and

tfp

at the Firm

Level: Do Intangibles Matter?

MICHELE BATTISTI

Department of Law, University of Palermo,

Piazza Bologni 8 - Palermo (Italy), CeLEG LUISS Guido Carli and RCEA

michele.battisti@unipa.it

FILIPPO BELLOC

Department of Economic Studies, “G. d’Annunzio” University,

Viale Pindaro 42 - Pescara (Italy)

f.belloc@unich.it

MASSIMO DEL GATTO

Department of Economic Studies, “G. d’Annunzio” University,

Viale Pindaro 42 - Pescara (Italy), and CRENoS

m.delgatto@unich.it

We use a panel of European firms to investigate the relationship between intangible assets and

productivity. We distinguish between total factor productivity (tfp) and technology adoption,

whereas standard estimations consider only a notion of productivity that conflates the two

effects. Although we are unable to address simultaneity, we allow for the existence of multiple

technologies within sectors through a mixture model approach. We find that intangible assets

have nonnegligible effects that both push firms toward better technologies (technology adoption

effects) and allow for more efficient exploitation of a given technology (tfp effects).

1. Introduction

By including all those assets that lack a physical dimension (i.e., quality of management,

customer loyalty, information infrastructure, trade secrets, research and development

(hereinafter R&D), and, more generally, a company’s intellectual capital), intangible as-

sets form the “knowledge base” of a firm and are often found to play an important role

in modern knowledge-intensive production (e.g., Delgado-Gom`

ez and Ramir`

ez-Ales`

on,

2004; Hall et al., 2005; Bontempi and Mairesse, 2008; O’Mahony and Vecchi, 2009). How-

ever, whereas the existence of a positive relationshipbetween intangible assets and firm

performance is now widely accepted (see, e.g., Oliner et al., 2008), empirical research on

the channels through which the relationship takes place is rather scant. Notwithstanding

a few contributions showing that investments in intangible assets foster productivity

at the firm level (Bontempi and Mairesse, 2008; O’Mahony and Vecchi, 2009; Marrocu

et al., 2012), it remains unclear to what extent such productivity gains occur throughtotal

factor productivity (hereinafter tfp—intended, in a strict sense, as the ability to exploit

“traditional” inputs) and through a process of technology upgrading that is induced by

an increased ability to identify and adopt more productive technologies. In particular,

Wethank seminar participants at the 2012 Workshop of the International Study Group on Export and Produc-

tivity (ISGEP) in Stockholm and the 2014 Geography of Innovation Conference at the University of Utrecht.

We also thank the co-editor and two anonymous refereesfor insightful comments.

C2015 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume24, Number 2, Summer 2015, 390–414

Unbundling Technology Adoption and tfp 391

extant productivity analysis is silent on whether intangible assets affect the process of

technology adoption because standard productivity measures are only able to consider

a notion of productivity that conflates technology and tfp.1

In this paper, we examine the relationship between intangible assets and produc-

tivity in a large sample of European manufacturing firms by adopting a tfp estimation

strategy that enables us to distinguish between technology effects and tfp effects. We do

so against the background of a world in which different production function coefficients

identify different technologies. Several technologies areavailable in each sector-industry,

with a number of firms using each technology. To aid in the definition of terms, let us

anticipate the formal description and consider the following production function:

Yi,t=Ai,t

N



n=1

(Xn,i,t)βn,m,(1)

where Ai,tis firm i’s tfp (i.e., “Solow residual”) at time t,Xn,i,tdenotes the amount

of input nused by firm iat time t,βn,mis the associated production coefficient, and

Yi,tdenotes produced output. Index mis introduced to refer to a specific “technology,”

with m=1,...,Mand Mdenoting the number of available technologies. According

to equation (1), the amount of output firm iis able to produce given the amount of

inputs depends on two factors: the first (Ai,t) is firm-specific, whereas the other (βn,m)

is intrinsic to the adopted technology and is common to all the firms that use the same

technology.A group of firms sharing a given technology may or may not share the same

industry. Whereas firms’ choices concerning these two factors are usually referred to as

technological choices or productivity choices, we reserve the term “technology” to refer

to βn,mand discuss tfp with regard to Ai,t.

Our analysis aims to examine whether the stock of intangible assets plays a role in

productivity by separating out the effect that occurs through technology adoption (i.e.,

through βn,m) from the effect that occurs through the tfp term Ai,t. We first employ mix-

ture models to estimate the production function parameters in equation (1), allowing for

the existence of two technologies within each sector; we then use a first-order stochastic

dominance criterion to identify the “high” and “low” technology. This method allows

us to cluster our sample firms over the two technology groups and compute the tfp com-

ponent in equation (1) for each firm as the difference between the actual and predicted

output, given the technology adoption. Finally,we estimate the effect of intangible assets

on the probability of belonging to the high technology group (technology adoption effects)

and on the ability to exploit the technology in use (tfp effects).

We find that intangible assets have positive and statistically significant effects

on both technology adoption and tfp. For the firms in the low-technology group, the

estimated increase in the probability of choosing the “high” technology associated with

a 1% increase in the intangible-to-tangible assets ratio yields an expected gain in value

added ranging from 0.89 to 1.56 on average. For the same increase in intangible assets,

the value added for all firms is augmented by 1.17% due to the tfp effect.

The paper proceeds as follows. In Section 2, we briefly describe how our anal-

ysis contributes to the economic literature. In Section 3, we apply mixture models to

production function estimation and obtain firms’ technology clusters and tfp values. In

Section 4, we separately measure the impact of intangible assets on technology adoption

1. To highlight this limitation, Bernard and Jones (1996) refer to such “mixed” measures of productivity

as “total technology productivity.”