Innovation, Tangible and Intangible 当 Investments and the Value of Spanish Firms by Aitor Lacuesta^**, Omar Licandro^***, Teresa Molina^**** and Luis A. Puch **** ***** Documento de Trabajo 2009-19

June 2009

We thank Juan Ramón García and Massimiliano Marinucci for excellent research assistance through different stages of this project. We also thank Jorge Durán, Elena Huergo, Juan Francisco Jimeno, Pedro Mendi, Teodosio Pérez, Dirk Pilat, José Antonio Moreno and Adelaida Sacristán for very helpful comments, as well as seminar audiences at Granada SAE 2007, Toulouse Knowledge Conference, and Barcelona Zvi Griliches Summer School. Part of this research was done while Puch was visiting the European University Institute under the Fernand Braudel fellowship programme which is gratefully acknowledged. Financial support from Fundación COTEC and the Fundación Focus-Abengoa, as well as the Dirección General de Investigación, project SEJ2007-65552, are gratefully acknowledged.

** Bank of Spain.

*** European University Institute.

**** World Bank and FEDEA.

***** Universidad Complutense and FEDEA.

Los Documentos de Trabajo se distribuyen gratuitamente a las Universidades e Instituciones de Investigación que lo solicitan. No obstante están disponibles en texto completo a través de Internet: http://www.fedea.es. These Working Paper are distributed free of charge to University Department and other Research Centres. They are also available through Internet: http://www.fedea.es.

ISSN:1696-750X

Jorge Juan, 46 28001 Madrid -España Tel.: +34 914 359 020 Fax: +34 915 779 575 infpub@fedea.es

Aitor Lacuesta^𝑎, Omar Licandro^𝑏, Teresa Molina^𝑐 and Luis A. Puch^𝑑†

^𝑎Bank of Spain ^𝑏European University Institute ^𝑐World Bank and FEDEA ^𝑑Universidad Complutense and FEDEA

April 2009

Abstract

Why is R&D spending so low in Spanish firms? One possible answer may lie in a small contribution of innovative investments to value creation at the firm level. When pulling together complementary sources of spending data and related evidence to measure these investments, we observe that R&D is low for international standards, but overall intangible investment seems adequate. Data from the Central de Balances are then used to assess the efect of R&D and other innovative investments on the value of Spanish firms. The results suggest that intangible investments have a positive impact on market values which is more substantial for innovative sectors, and this is also the case for R&D capital. Such a positive impact is influenced by the size of the firm and its presence in the stock market. In fact, an alternative explanation to low R&D intensity could be found in the small fraction of firms publicly traded in the stock market in Spain, as far as equity holders tend to value intangible assets more than bond holders. Consequently, promoting a more active role of market valuations as a guide for innovative investment might be a promising policy.

Keywords: Tangible Investment, Intangible Investment, Market Value JEL Classification: E22, C33, L60

^∗We thank Juan Ramón García and Massimiliano Marinucci for excellent research assistance through diferent stages of this project. We also thank Jorge Durán, Elena Huergo, Juan Francisco Jimeno, Pedro Mendi, Teodosio Pérez, Dirk Pilat, Jose Antonio Moreno and Adelaida Sacristán for very helpful comments, as well as seminar audiences at Granada SAE 2007, Toulouse Knowledge Conference, and Barcelona Zvi Griliches Summer School. Part of this research was done while Puch was visiting the European University Institute under the Fernand Braudel fellowship programme which is gratefully acknowledged. Financial support from Fundación COTEC and the Fundación Focus-Abengoa, as well as the Dirección General de Investigación, project SEJ2007-65552, are gratefully acknowledged.

^†Corresponding Author: Luis A. Puch, FEDEA, Jorge Juan 46, 28001 Madrid, Spain; E-mail: lpuch@fedea.es

1 Introduction

Innovation is a complex process involving both creation and adoption. These activities require investments in human and physical capital, together with some other forms of intangible assets. Indeed, not all investments are innovative. However, those activities accumulating in knowledge allow for the progress of Total Factor Productivity (TFP) and the creation of value in the firm.

The question of how investment in diferent types of capital brings about innovation, and how this process afects the performance of the economy is of considerable importance. A large academic literature has built upon the sources-of-growth framework developed by Solow (1957, 1960) and others, in the 1950s and 1960s, for sorting out the factors that drive Total Factor Productivity (TFP) growth. Advances in this framework have lead to questions related to the emphasis on quality change in the measurement of prices, or to the efects of capitalized intangible investments. As for the latter, from the theoretical side, the intangible capital extension of neoclassical growth theory has been used to measure the amount of intangible assets in the US economy (McGrattan and Prescott (2006)). Alternatively, from the empirical side, a number of recent papers have measured business spending on intangible capital (Corrado, Hulten and Sichel (2005)) and quantified its role relative to the various types of produced capital, by estimating production functions (see Doraszelskiy and Jaumandreu (2007), and the references therein) and market value equations (Brynjolfsson and Hitt (2002) for the US, or Hall and Oriani (2006) for the US and some EU economies).¹

In this paper, we retain the empirical approach to measure tangible and intangible assets and to quantify its efect on the market value of Spanish firms, notably the impact of R&D activity. We proceed in two steps. First, we build a measure of tangible and intangible investment in physical and human capital in line with Corrado et al. (2005). To this purpose, we report measured National Accounts investment data and we pull together complementary pieces of spending data on intangible assets (Encuesta de Innovación Tecnológica (EIT),

¹An OECD panel aims at improving estimates of the scale of investment in intangible assets at the national level for selected countries (Finland, Japan, Netherlands, United Kingdom and United States) based upon such an intangibles measurement approach. As for production function estimation, as early as Griliches (1979) proposed to augment the production function with the stock of knowledge capital, as proxied by a firm’s accumulated R&D expenditures.

Panel de Innovación Tecnológica (PITEC), and other related evidence). This measure gives insight on the diferent investment intensities and its evolution at the aggregate level. Second, we estimate the impact of measured intangible investments relative to tangible investments and proxies for unmeasured intangible assets from the firm-level data collected in the Central de Balances of the Bank of Spain following the market value approach. The fact that the Central de Balances (Firm’s Balance Sheets) database is an incomplete census of the nonfinancial business sector in Spain precludes its use to build a measure of aggregate spending on tangible and intangible assets. Therefore, we draw inference from the micro estimates to the patterns of aggregate investment to ask whether the low R&D intensity observed in the Spanish economy has to do with a small contribution of R&D capital to the value of the firm.

Several authors have explored the economic consequences of diferent intensities in intangible capital investment relative to tangible investment either among firms or countries. Notably, the question of how large is the stock of intangible capital is crucial to properly measure productivity during times of changing investment, as it has been the case of the last two decades. Abowd et al. (2005) and Black and Lynch (2005) discuss how intangible assets, and notably R&D capital, contribute to explaining diferences in productivity among firms. At the aggregate level Parente and Prescott (2000) show that intangible capital is needed to account for diferences in per capita income among countries. Also, intangible capital has been considered of much interest to judge whether the stock market is correctly valued (McGrattan and Prescott (2005)). Another related approach has used the market value of the firm as an indicator of the economic results from investing in knowledge capital [cf. B. Hall (1993)] or to infer the product of capital in relation to the quantity of intangible capital. For instance, R. E. Hall (2001) discusses how increases in the ratio of stock-market value to capital are associated with high values of the product of capital. Unmeasured intangible capital is a residual after subtracting the capitalized value of measured tangible and intangible assets from the total value of corporations.

As these authors we relate tangible and intangible investment to TFP growth or market values. Diferent from them we stress on a broad description of the input of innovative activities related to the creation and adoption processes. In particular, we ask whether innovative investments have a stronger impact on market values, and the database we use is unique for this purpose. Behind the argument is the idea that technical progress is embodied in new equipment or new vintages of human capital. Corrado et al. (2005) label expenditures on intangibles as knowledge capital in the sense that it is well represented by intangible capital accumulation. This knowledge often takes the particular form of innovative investments in R&D. Whether TFP growth or value creation is a function of “fundamental” research or total business spending in intangibles is mostly a measurement issue, possibly sector-specific.² Further, R&D is technology, not capital. Consequently, it is dificult to interpret what does a low investment intensity or a small stock of R&D capital mean.³

To provide further insight on the size and extent of R&D investment, aggregate data from various surveys are used to measure R&D expenditures as well as tangible and intangible investments in physical and human capital. We find that R&D investment is low for international standards, whereas overall intangible investment turns out to be of the right magnitude. While it is true that R&D is exhibiting the higher growth rates in recent years among the diferent categories of intangible investments, this enhanced activity is not enough to compensate the initial unbalance. The question is then whether the Spanish data of the financial statements of firms in the Central de Balances can help to interpret this quantity anomaly. The finding is that the market value of non-financial firms in Spain turns out to be meaningfully related to their knowledge assets, and in particular with R&D capital. This link is stronger the more technological is the sector the firm belongs to. Therefore, a low R&D intensity in the Spanish economy does not come from a small return of this form of investment. However, an issue to be addressed is that the impact of R&D activities exhibit a high correlation with size and the firm’s presence in the stock market. The small number of publicly traded firms in our sample precludes a deeper look into the role exerted by public stock markets. Nevertheless, the fact that intangible assets are more valued by equity holders than by bond holders, and its implications for R&D intensity, is being explored in Lacuesta et al. (2008).

The paper is organized as follows. Section 2 examines the aggregate investment position of the Spanish economy in connection with innovative activities, and provides estimates for the various sources of tangible and intangible capital. Section 3 proceeds with an econometric evaluation of the creation of value at the firm level of alternative investments. Each section discusses its corresponding theoretical background. Section 4 draws inference from the micro evidence obtained to asses innovation in the Spanish economy from its aggregate investment position. This section includes as well some concluding remarks.

²In Licandro and Puch (2008) an empirical implementation of these methods to the energy sector is discussed, with application to the subsector of renewable energies as a case study.

³Compared to capital, technology or knowledge capital embeds an invention cost (once per technology) and an adoption cost (once per user), beyond these there are only user costs [cf. Jovanovic (1997)].

2 The Inputs for Innovation in the Spanish Economy

2.1 Innovation and Investment

Innovation is a complex phenomenon. To characterize innovation we build upon the operative definitions reported in oficial statistics. For instance, the Community Innovation Survey (CIS, EUROSTAT) reports data on product innovation, process innovation and organizational innovation, as well as more recently on marketing innovation. The following EU-wide definitions of the Oslo Manual can be used. Thus, a product innovation is the market introduction of a new or a significantly improved good or service. A process innovation is the implementation of a new or a significantly improved production process, distribution method or support activity for the firm’s good or services. In both cases it must be new to the firm not to the sector or market, and it does not matter if it was originally developed by the firm or by other firms, that is created or adopted. On the other hand, an organizational innovation is the implementation of new or significant changes in firm structure or management methods intended to improve firm’s use of knowledge or the eficiency of work flows. A marketing innovation is the implementation of a new or a significantly improved designs or sales methods. Again, in both cases it must represent an increase in the quality or appeal of firm’s goods and services.

Clearly, all these environments for innovation can be related to diferent forms of tangible and intangible investments. Indeed, not all investments are innovative. Our strategy here is collecting all forms of investment and see whether some forms of investment have a greater share than others at the aggregate level. Then we will explore whether those quantity diferences if any can be justified by diferences in the return of the diferent investments at the firm-level. Since we will find a quantity anomaly between R&D and other forms of intangible investment, and that these investments have a larger impact on market values in innovative sectors, we will concentrate on the size and impact of innovative investments.

We briefly discuss first the theoretical framework underlying this approach, and then we proceed in detail with a broad measurement of recent investment in the Spanish economy. The findings below will motivate the estimation of the market value–tangible/intangible capital relationship in the following section.

2.2 Theoretical background: growth models and factor accumulation

The theoretical background we consider justifies the importance of a broad measurement of investment, and the importance of investment under the hypothesis of Embodied Technical Progress (ETP). Prior to the idea of ETP, we start from neoclassical growth theory (`a la Solow) to arrive to the consideration of knowledge(-varieties) as the way to endogenize technical progress.

In a seminal contribution, Robert M. Solow (1957) suggests a procedure to measure the contribution to economic growth of factor accumulation and technical progress, in a consistent way with Neoclassical growth theory. Aggregate output is assumed to be produced according to a Cobb-Douglas technology

\[Y _ {t} = A _ {t} K _ {t} ^ {\alpha} L _ {t} ^ {1 - \alpha},\tag{1}\]

where is output, and are capital and labor allocated to production, respectively, and is the state of technology, the so-called total factor productivity (TFP). Consequently, the growth rate of technical progress can be measured as

\[g _ {A} = g _ {y} - \alpha g _ {k}.\tag{2}\]

where is the growth rate of output per worker (or hour worked) and is the growth rate of the per worker stock of capital. Observed measures of GDP and the labor input, as well as constructed measures of the capital stock are usually used to identify technical progress. In order to measure the capital stock, the Neoclassical theory assumes that output is allocated to consumption and investment according to

\[Y _ {t} = C _ {t} + I _ {t},\]

and capital accumulates following the method of permanent inventories

\[K _ {t + 1} = (1 - \delta) K _ {t} + I _ {t},\]

where past capital depreciates at the rate 𝛿, . Data on investment from national accounts are then used to build capital series.

The key issue is the proper measurement, as well as a proper understanding of TFP growth. The Solow procedure correctly identifies the sources of growth if output and inputs are correctly measured, the technology is Cobb-Douglas and the measure is exogenous to the time and place.

Attempts to endogenize the rate of technical progress relate to considering several forms of technology accumulation underlying the process. Approaches that have been used to this purpose consider external efects and knowledge spillovers. Also there are several extensions to two-sector models that exploit the existence of either more than one accumulable factor (R&D, human capital, intangible capital) or more than a source of technical progress (embodied or disembodied).⁴

According to Romer (1990), for instance, R&D activities accumulating in knowledge allow for the progress of TFP.⁵ Output from (1) can be allocated to consumption, , the accumulation of capital, and R&D investment, , according to

\[Y _ {t} = C _ {t} + I _ {t} + X _ {t}.\]

⁴Rather than accounting for technology accumulation, an alternative route leads to growth models with TFP diferences. This later theory is possibly needed to account for diferences in international income (see Parente and Prescott (2000))

⁵Contrary to models of external efects, growth is not a side product of factor accumulation, but is a result of profit-maximizing firms’ intentionally investing in R&D. In this framework, to make R&D activities profitable, researchers have the right to patent their discoveries, which give them some monopoly power.

The R&D technology creating new goods or improving the quality of existing goods makes TFP grow according to

\[A _ {t + 1} - A _ {t} = b \frac {X _ {t}}{Y _ {t}},\tag{3}\]

where parameter gives the marginal productivity of R&D production. The larger the fraction of resources allocated to R&D is, the large is technical progress and growth. In this setting, TFP growth depends on intangible capital accumulation, which takes the particular form of innovative investments in R&D. Whether TFP, is a function of “fundamental” research or total business spending in intangibles is mostly a measurement issue, possibly sector-specific.

One fundamental problem with the Neoclassical and the endogenous growth models is that they are inconsistent with the observation of a secular decline in the relative price of investment goods. The model can be extended so that it matches this secular decline by introducing an equipment investment good sector and assuming that technological change is faster in this sector (see Grenwood et al. (1997)). Therefore, in the framework of the Neoclassical growth model (intangible investment 𝑋 may be easily added to the analysis), let us assume that

\[C _ {t} + Z _ {t} = B _ {t} \hat {K} _ {t} ^ {\alpha} (L _ {t}) ^ {1 - \alpha},\tag{4}\]

where now represents the state of technology in the consumption goods sector, call it neutral (or disembodied) technical progress, and is a measure of efective capital whose meaning will become clear below. Diferent than the Neoclassical model, is not investment but an input in the technology to produce equipment according to

\[I _ {t} = q _ {t} Z _ {t},\tag{5}\]

where is a measure of technical progress specific to the investment goods sector, that is a measure of embodied technical progress, and is the flow that builds efective capital

according with

\[\hat {K} _ {t + 1} = (1 - \delta) \hat {K} _ {t} + I _ {t}.\]

In this setting, an economy that does not invest (in equipment) does not get the rewards from this second form of technical progress. This economy produces then two diferent final outputs, consumption and investment, using two diferent technologies with diferent rates of technical progress, that we denote and . The growth rate of technical progress, denoted as before, is a linear combination of the growth rates of both sectors, weighted by their contribution to total output,

\[g _ {A} = (1 - s) g _ {B} + s g _ {q}.\tag{6}\]

where the share of consumption and investment in output are denoted and respectively.

It is in this framework that the construction of a broad measure of investment is pursued, and assessed to account for the importance of diferent types of capital for the creation of value at the firm-level.

2.3 A broad measurement of investment

Investment can be defined as any allocation of resources designed to increase future production possibilities. Among tangible investment, new structures and equipment together with the changes in inventories are measured, whereas resources devoted to maintenance and repair go expensed. As discussed above, process innovation is notably related with equipment investment, and the embodiment hypothesis provides a rationale for this. Among the sources for Intangible investment we have R&D and Software that, if internal to the firm, either go expensed or they are measured as an intermediate input. In addition, organizational capital goes typically unmeasured, as it is the case with the most part of human capital accumulated within the firm and sweat equity. Product innovation and organizational innovation (see Black and Lynch (2005)), but also some forms of process innovation are closely related to all these forms of investment.

(Private Business) Fixed Investment

The National Accounts of the Spanish economy classifies Private Business Fixed Investment in Tangible Fixed Assets, as Structures (residential and non-residential), Equipment and Cultivated Land and Cottage, and Intangible Fixed Assets, as Computer Software, Mineral Exploration and Creative (leisure, literary, arts) Property, and other. In addition, (Big) Repairs of (non-produced) tangible assets (land) and Transfers for (non-produced) tangible assets (land, patents) are also measured as fixed assets.

Table 1 reports fixed tangible investment according to the categories collected in National Accounts and with special emphasis on the equipment investment figures. The top panel reports the figures in levels, and the bottom panel in percentage of GDP.

[INSERT TABLE 1 ABOUT HERE]

Fixed Investment

Table 2 reports fixed intangible investment according to the categories collected in National Accounts in levels and in percentage of GDP.

[INSERT TABLE 2 ABOUT HERE]

Intangible Investment

The value added by reporting these figures is small. The reported results serve to provide a broad description of measured investment. The more relevant findings can be summarized as follows. Gross Fixed Investment in the Spanish economy has moved from about 20% of GDP in 1995 to above 30% in 2005. It has been reported elsewhere that on average it was about 22% along the period 1964-95 (cf Estrada et al. (1997)). This remarkable increase observed in recent decades comes mostly from residential investment which has moved from 4% to nearly 10% of GDP. Over the recent period equipment investment has remained relatively constant at about 8% of GDP (in the manufacturing sector this ratio is 20%), and it has exhibited a nominal annual growth rate at about a 3.5%. Investment in other fixed intangible assets registered in National Accounts represents about 4% of fixed investment, that is, nearly 1% of GDP. This intangible investment figures mainly correspond to software acquisitions, and represent a small fraction of intangible assets in the Spanish economy as we document below.

(Private Business) Spending on Intangibles

To asses the economic significance of the fixed investment figures just reported, and to give a scope for how much of investment goes unmeasured or is not taken as such (goes expensed) in the Spanish economy, we build an estimate for the values of intangible investments in the Spanish economy from several sources. We follow Corrado, Hulten and Sichel (2005) in characterizing intangible investment along the following three categories: Computerized Information, Scientific and Creative Property and Innovative Property, and Economic Competencies. Next we briefly revise the content of these categories. Some further details can be found in Corrado et al. (2005), in its application to US data, and in Licandro and Puch (2008) in the Spanish case.

∙ Computerized Information

Computerized Information is defined as the knowledge capital embodied in computer software and computerized databases. More precisely, it is organized between

1. Computer Software

Among intangible computer software, the estimated costs of software created by firms for their own use is considered. Software purchased external to the firm it is already considered as fixed investment in the Spanish National Accounts (NA – SAE 93: CNAE95).

2. Computerized Databases

Spending in computerized databases refers to subscription costs to external or customized databases. These costs are not capitalized in NA, and typically represent a small figure. In the Encuesta Nacional de Servicios (EAS, 1998-2005), the sector of IT services is considered in detail (sections 4.1 and 5.15).

Corrado et al. (2005) call attention on the possible double-counting of software, and the overlap between figures for own-account software and the data on R&D expenditures.

Our strategy here is to report software expenditures as belonging to the measure of internal R&D that it is examined inside the next block (categories 3., 4. and 6. below). Therefore, Table 3 reports only (for completeness) our estimates for the volume of Computerized Databases spending as documented in the EAS, 1998-2005, both in levels and in ratios over GDP.

[INSERT TABLE 3 ABOUT HERE]

Computerized Information

The results summarized in the table are not very informative at this stage since we are leaving the software block apart, included in the R&D data. The value of Computerized Databases has remained fairly stable at about 0.2% of GDP, slightly increasing over the period though.

∙ Scientific and Creative Property, and Innovative Property

This block includes scientific and nonscientific R&D, the second category related to artistic and innovative knowledge embedded in commercial copyrights, licenses and designs. This second category is not very well measured, in contrast to the scientific component embedded in patent, licenses and general know-how (not patented).

3. Scientific and Engineering R&D

This category includes R&D expenditures in manufacturing, as well as those in the software and the ICT industries. We approximate these figures building upon the data reported in the Estadística sobre Actividades de I+D (EI+D), Encuesta de Innovación Tecnológica (EIT) and Indicadores de Alta Tecnología (CNAE93 section D–manufacturing, and industries 64(2)–ICT, and 72(2)–software).

4. R&D in Finance and other Services

R&D expenditures in the financial sector and other services sectors. We approximate these figures from the EIT and PITEC data (sections J and K –financial, sectors 742–services in architecture engineering, and 743–technical services, and the part of R&D in social sciences, division 72).

5. Mineral Exploration

R&D expenditures in mining, collected again from the EI+D and the EIT (section C, sectors 10 to 14 and estimates for subsectors 45112, 7420(3)-(4)–land exploration and topography services).

6. Copyright and License Costs

With category 4. above, this one completes the nonscientific R&D. Here, R&D in the literary and artistic edition industry as well as the broadcasting industries. We deduct this part from category 3. (edition industries from division 22, plus sectors 921 and 922 and additional information collected in the EAS, formerly– 1997– actually the Survey of Broadcasting Services).

Table 4 reports our estimates for the volume of Scientific and Creative Property, and Innovative Property in the Spanish economy, in levels, whereas Table 5 reports the figures in ratios over GDP.

[INSERT TABLE 4 ABOUT HERE]

Scientific and Creative Property, and Innovative Property (Levels)

Total spending among these categories doubled from 1998 to 2005, which implies a nominal growth rate of these spending close to 10% over these years. We complete the EIT data with recent data obtained from PITEC, for which we associate the fraction of investment surveyed in the EIT that the panel extract in the PITEC is retaining. The overall picture is close to the one for R&D in Science and Engineering. The rest of the categories are only recently available and in many cases are only considered in the PITEC.

[INSERT TABLE 5 ABOUT HERE]

Scientific and Creative Property, and Innovative Property (Ratios over GDP)

Total spending along this category, which is somewhat an augmented R&D record for industries and services other than the R&D producing industry, reaches nearly a 1 per cent of GDP. This number roughly coincides with the total R&D figures for Spain, so what it is added here diferent from R&D, it is compensated with what it is left out because already measured in value added. It turns out that this figure is in a ratio 1 to 5(-8) to the comparable estimate in Corrado et al. (2005) for 1998-2000 in the US. This result is hardly explained by the relative position of the Spanish economy to the US itself. We could roughly characterize this position in relative terms (so leaving apart the US is 10 times bigger) by two features: a) the Spanish economy is 70 per cent as rich in income per capita (at international prices), b) this results on average in a one half technology intensity compared to the US. This two elements would justify, say, a 1 to 3 comparison with respect to the US.

∙ Economic Competencies

This category is possibly the most dificult to measure and includes brand equity, firmspecific human capital and organizational capital. The total for these categories is above 2 per cent of GDP during the 2000-2005 period.

7. Brand Equity

Investments in this category include expenditures on advertising and market research. This covers the costs for launching new products, the creation of customers’ lists, and the maintenance of the value of brand names. Both the EIT and the PITEC collect these data. There is also relevant information in the EAS in this respect, and there is the data from the Encuesta sobre Estrategias Empresariales (ESEE) that could be used to complement these figures.

8. Firm-specific Human Capital

This category includes the costs of developing workforce skills. EIT and PITEC collect these data when corresponds to workers directly engaged in innovative activities. Again, for a more general estimate the ESEE data could be also used here.

9. Organizational Structure

For an estimate of investments in the organizational structure information on executives’ wages and management consulting fees are used. Both of these data are collected in the EAS, section 5.20.

[INSERT TABLE 4 ABOUT HERE]

Economic Competencies (Levels)

[INSERT TABLE 5 ABOUT HERE]

Economic Competencies (Ratios over GDP)

This last category constitutes a traditionally important component of intangible investment. In levels it has been steadily growing since 2000. However, its share of GDP seems to have been downsizing mostly along the brand equity and developing workforce skills categories, and less with respect to organizational capital. Of course, these numbers are possibly the more subjected to further revision.

Summarizing, there is a lot of value added in the intangibles block of our exercise. First, we have searched for the sources of these data in the Spanish economy. Second, we have selected the more reliable data and we have put them together in a comprehensive presentation. Finally, the quantities reported are of interest for many related applications. The main results can be summarized as follows. Business spending on intangibles in Spain was about EUR 22 billions in 2000, and above EUR 25 billions in recent years annually, to reach up to 3.5 per cent of GDP. The absolute figure can be compared to an aggregate level for the US of about $1.2 trillion (cf. OECD (2008) as well) which can be obtained from the Spanish one , with these factors corresponding to size, income and technological diferences, respectively. In particular, the latter factor relates to Eaton and Kortum’s (1999) idea of international technology difusion, and the diferences in the contribution to the technological frontier of some countries with respect to others. The intensity figure can be compared with a 13 per cent in the US, and relative to other Spanish data with the 7 to 9 per cent of GDP in machinery, and the 0.7 to 0.9 per cent in R&D. Since 2000, and to 2005 we do not observe (yet) major changes in this ratio. One question is whether there are any major changes to be detected soon. Scientific and Creative Property (R&D) items seem to have increased at rates above 10 per cent annually, to represent from about one fourth of the estimate for Economic Competencies in 2000, to reach nearly one half of this traditionally important component of intangible investment. However, the R&D figures remain still in a ratio, say, one to five relative to the comparable figure for the US, whereas the total spending over GDP for intangible investment is closer to a more interpretable 1 to 3 ratio with respect to the estimates reported by Corrado et al. (2005) for the US. One may wonder whether apart from the process of creation of innovations (mostly R&D spending), technological progress requires the same relative spending in intangibles, so far as the economy is mostly involved in the process of adoption of innovations.

3 Innovation and Value Creation at the Firm Level

3.1 Theoretical background and the empirical model

The theoretical framework for market value analysis can be derived from a standard dynamic optimization problem for the firm. The value of the firm at time 𝑡, 𝑉, can be expressed in the form of Bellman equation according to

\[V (A, K) = \max _ {K ^ {\prime}} \Pi (A, K) + \beta E _ {A ^ {\prime} | A} V (A ^ {\prime}, K ^ {\prime})\]

where 𝐴 is the state of technology, 𝐾 is the aggregate stock of capital that evolves to build next period capital stock from the flow of investment, and the undepreciated part of the diferent types of capital, according to

\[K ^ {\prime} = I + \sum_ {j} \omega_ {j} k _ {j}, \text { all } j\]

and Π is the instantaneous flow of profits at 𝑡

\[\Pi (A, K) = A F (K, N) - W N - I\]

where 𝑁 is labor and 𝑊 the real wage rate. With constant returns to scale, all assets documented, and no adjustment costs, buying a firm is equivalent to buying a collection of separate assets. Therefore, the market value of the firm can be expressed as an additive

function of its single assets (cf. Hall (1993))

\[V (A, K) = \sum_ {j = 1} ^ {J} k _ {j},\tag{7}\]

for 𝐴 being set to its unconditional mean. Further, whenever a vintage 𝑧 plant (a plant of age can be described as

\[{Q _ {t} (\tau)} {= A _ {t} (\tau) F (\kappa_ {t} (\tau), N _ {t} (\tau))}\]

where

\[\begin{array}{r c l} {I (t)} & = & {\kappa_ {t + 1} (0)} \\ & & \\ {\kappa_ {t + 1} (\tau)} & = & {(1 - \delta) \kappa_ {t + 1} (\tau - 1).} \end{array}\]

With constant returns to scale, all plants with the same Total Factor Productivity at 𝑡 and the same number of plants for all 𝑡 (cf. Baily (1981)),

\[V (A, K) = \sum_ {\tau = 0} ^ {\infty} q _ {t - \tau} \kappa_ {t} (\tau) \equiv K\tag{8}\]

for 𝐴 being set to its unconditional mean. Therefore, under standard regularity assumptions (zero adjustment costs) and perfect capital markets either heterogeneity by types of capital or by ages of capital are aggregated out.

and Econometric Model

Theory (market value equations (7) or (8)) suggests a basic estimating relation between the market value of firm 𝑖 and the assets the firm possesses, allowing for repeated observations over time 𝑡

\[V _ {i t} = \alpha_ {i} + \sum_ {j = 1} ^ {J} \nu_ {j} K _ {j, i t} + \varepsilon_ {i t} ^ {\nu}\tag{9}\]

We would expect all 𝑖 and under the ideal conditions for market value equations (7) or (8). However, ¯𝜈 deviates from 1 if adjustment costs are present, or omitted variables are correlated with observed assets. We follow Brynjolfsson et al. (2002) in interpreting as the diference in value between type of capital installed into the firm and otherwise identical capital available in the market. Descriptive results for our sample in this stylized theoretical framework are discussed below. We also consider additional covariates, and we deal with firm and time-specific efects to address factors specific to individual firms. Further econometric issues as sample-selection bias or the problem of omitted variables are preliminary discussed in Lacuesta et al. (2008). Finally, we follow Hall and Oriani (2006) in an alternative formulation of equation (9) that makes explicit that the ¯𝜈 coeficients are not structural parameters but equilibrium outcomes in the market at time 𝑡, and therefore, they has to be interpreted as a measure of the current average marginal shadow value of an additional currency unit spent on capital asset

Data

All the accounting data of firms that we use in this part of the analysis come from the Central de Balances of the Bank of Spain. This database includes financial statements for an unbalanced panel of non-financial Spanish corporations (consolidated accounts from firms) which is available at the Bank of Spain. Our sample consists of 2872 traded and non-traded firms and 20575 observations over the period 1992-2005. Further details on the selection of the sample can be found in Lacuesta et al (2008). In particular, it was in 1992 that a reduced questionnaire was introduced, together with a new, more homogeneous accounting regime. Also, further descriptive statistics of the sample, with applications to the energy sector, are presented in Licandro and Puch (2008).

The database includes the treatment of the asset position (capitalization) of the firm according to well established accounting principles. To the book value of tangible and intangible assets the Bank of Spain associates a record of the market value for traded firms, together with an estimate of the value of equity and other shares for non-traded firms (see B de E 2005). This makes the selected database a unique source of data for market value analysis.

Our dependent variable is the market value either reported, for traded firms, or estimated for non-traded firms by the Bank of Spain. The explanatory variables are the book value of diferent asset that conform the total asset position reported by firms. Consequently, our quantitative results are based on raw data without further refinements. The kind of refinements needed to make available the corresponding data for several countries are thoroughly discussed for instance in Hall and Oriani (2006). These refinements could be adapted to our sample as a robustness check. In particular, we follow closely the strategy proposed by these authors in examining the market valuation of diferent intangible assets, as discussed below.

3.2 Key Results: Market Value and Asset Quantities

We regress market value on book values for the aggregates of tangible and intangible assets. We do this for stock market, quoted and non-quoted corporations. We also proceed with a finer disaggregation of those assets reported in the balance sheet at the Central de Balances level.

There are several econometric problems to be addressed with these data. Among them, we mostly deal with firm and time-specific efects. We also instrument the explanatory variables to test for endogeneity. Sample-selection bias, the problem of omitted variables and other econometric issues are further discussed in Lacuesta et al. (2008).

Investment and Market Value: Whole Sample

Table 8 reports results of regression analysis for the relationship between diferent types of assets and market value, all variables are in nominal terms, in millions of Euros. This regression includes physical tangible assets (equipment and structures), intangible assets (principally, capitalized R&D expenditures, industrial property and other intangible assets) and computer assets. Software is treated independently below, since it is only available after 2001. We also include measures of labor productivity and debt to assets ratios as controls whose coeficients are reported if they are significant.

[INSERT TABLE 8 ABOUT HERE]

Market values on Asset Quantities

The first column in Table 8 reports the results when tangible and intangible assets are the only explanatory variables. According to these, each euro of value of tangible assets explains EUR1.2 of market value. On average, intangible assets explain five times more of the market value of the firm than the value of tangible capital. The second column shows further that a major component of tangible capital are the computer assets. When computer assets are treated independently from the rest of tangible assets, each euro of computer assets is associated with about EUR15 of market value. This apparent excess sensitivity of market value to computer assets may suggest the presence of adjustment costs or other omitted components of market value correlated with these assets.

To more precisely account for the valuation of computer assets we consider a measure of other assets, computed as the diference between the book value of total assets and the sum of tangibles and intangibles. Column (3) attributes about EUR1.15 of market valuation to each euro of these additional assets, with the contribution of both tangible and intangible assets to value being slightly qualified with respect to the base case (1). Taking these additional assets into account (column (4)) we find that the market value of computer assets is close to EUR10, whereas the coeficient of other assets component is essentially the same. The result that the market valuation of intangible assets is five times bigger than that of equipment and structures is quite stable across specifications.

The finding that the coeficient for computers is about ten, whereas other types of capital receive coeficients below one, does not reflect that investment in computers earns an excess return. On the contrary this finding reveals a strong correlation between the stock of computer assets and unmeasured and much larger stocks of intangible assets.

The role of the controls is limited or non-significant across regressions. Because we are pooling multiple firms in multiple years, in all of the cases we include dummies for each year and sector at the second digit. Further, time and industry controls are jointly significant, so we are able to remove temporal shocks and omitted components relative to time period and industry. Note however that the coeficients for 2004 and 2005 year dummies are significant and sizeable (not shown). We use robust standard errors that are reported in parentheses.

Indeed, various types of firm-specific assets or organizational practices that are timeinvariant can contribute strongly to the market value of the firm. One way to account for these assets is a fixed efects (FE) estimation of the market value equation. Column (5) in

Table 8 suggests that once we remove the contribution of any time-invariant, firm-specific component of market value the valuation of computer assets is essentially the same, but now significant only at 5%. On the other hand, the variability in market value seems to remain significantly explained, and closely in a one to one ratio, by the measure of other assets.

The result obtained in the pooled sample that each euro in intangible assets contributes to market valuation five times more than each euro in non-computer tangible assets is nonsignificant under fixed efects. In this respect, it is worth noting that market values typically capitalize long-run characteristics of firms. Therefore, conditional on fixed efects, it is not surprising that annual realizations of tangible and intangible assets do not have a significant role in afecting market values. This turns out to be the case once computer assets are introduced separately into the analysis, and as far as the coeficient for these assets has a sizeable valuation.

One way to account for time-invariant firm-specific component of market value over moderately long periods is by estimating a diference specification. Varying diference lengths allows comparison of short-run (pool) and long-run (fixed efects) relationships. Table 9 reports estimates for the benchmark case (4) for intermediate diference specifications. The relationship for intermediate diferences seems substantial for computer assets, and it is somewhat preserved with respect to the pool for intangible and other assets. We conclude that it is the efect of tangible assets on market valuation what it is less precisely measured in the whole sample.

[INSERT TABLE 9 ABOUT HERE]

Market Values on Asset Quantities: Long Diferences

Therefore, ordinary least squares estimates of the pooled sample seem to be robust to long diferences. Also, model incorporating firm-specific fixed efects essentially preserves the significant impact of computer assets and the explanatory role of the variability in other assets. Finally, the results are robust to instrumental variable estimation over lagged explanatory variables. Whereas the variable lagged intangible assets is non-significant, last column in Table 8 reports two-stage least squares estimates when the variable lagged other assets is incorporated. In this case, the results are essentially preserved. Moreover, the impact of tangible and intangible assets seems more precisely measured when compared to fixed-efects estimation. Some other robustness checks are available upon request.

\[[ \text { INSERT TABLE 10 ABOUT HERE } ]\]

Market Values on Asset Quantities, by Sector of Activity

Finally, we explore what are the diferences between sectors for the benchmark regression equation (4). In Table 10 sectors are ordered from those for which a higher weight of intangible assets could be expected to those one might expect a stronger link of tangible assets to market values. The results clearly confirm the economic intuition and investment patterns are ordered across sectors. The second column retains the result for the whole sample. The label High Technology captures a broad sample of high and medium technology sectors. Again, a strong correlation between the stock of computer assets and unmeasured assets is found for this group of sectors. The energy sector for instance exhibits the higher estimate for the value of tangible assets contribution to market values. However, the way it compares with the building sector is very much in favor of the energy sector in terms of the role of intangible values, and thus of innovative activity according to our interpretation. The results under fixed-efects by sectors can be interpreted in line with the discussion above for the pool. Next we proceed with a finer disaggregation of assets.

Investment and the Market Value of R&D

As indicated above, finer disaggregation of intangible assets is possible since 2001. This disaggregation involves capitalized R&D expenditures, industrial property (patents, good will) and software expenditures. Here, we build upon the previous results to focus on R&D capital. To this purpose we follow Hall and Oriani (2006) in expressing all the alternative assets relative to the tangible assets of the firm. This strategy allows us to account for the relative size of intangible assets at the firm level, at the same time we disregard the asset component whose valuation seems to have a more limited impact in the value of the firm. More precisely, according to these authors, a version of econometric model (9) can be rewritten as

\[l o g (V _ {i t} / A _ {i t}) = l o g q _ {t} + l o g (1 + \gamma K _ {i t} / A _ {i t} + \lambda I _ {i t} / A _ {i t}) + \varphi_ {i t}\tag{10}\]

where the ratio is a proxy for average Tobin’s the ratio of the market value of the firm to the book value of tangible assets, and thus reflects the average market valuation of a firm’s total assets. 𝐾 is the value of R&D capital and I is the value of other intangible assets and consequently, and are the relative shadow values of these assets to tangible assets. With the approximation we can estimate (10) by OLS, in a specification closer to model (9) which is in levels and assumes no adjustment costs. Otherwise we report estimates by NLLS, and the corresponding slope coeficients comparable with the OLS estimates.

We perform regressions of market values on asset quantities for the diferent categories of intangible assets. The more robust results are obtained when we examine the market value of the firm as an indicator of the firm’s expected result from investing in R&D. Notice that we restrict the sample to those firms declaring to be engaged in R&D activities. Hall and Oriani (2006) perform separate regressions for each group of firms and they find that R&D disclosure is closely related to firm size. Table 11 reports estimates for regressions of market value to tangible assets ratio on the corresponding ratios for R&D on the one hand, and other intangible assets on the other. The regressions controls for the log of sales. The first column reports the OLS estimates for these assets, and suggests that each euro invested in R&D contributes to market valuation of the firm about three times more than the rest of intangible assets. Further, the coeficient is not significantly diferent from the equilibrium value of unity. These results are in line with those reported for the UK by Hall and Oriani (2006), which represents the country case most favorable for market valuation of R&D capital among the countries these authors consider: US, UK, France, Germany and Italy.

[INSERT TABLE 11 ABOUT HERE]

Market Values on Asset Quantities: R&D

The next two columns report the results of the NLLS estimation. The relevant parameter is now the coeficient divided by one plus the weighted average of the capitals. The result of computing that ratio at the variable is shown below the corresponding estimate, and the result of averaging the estimated coeficient for each firm is shown below that. These values are lower than both the OLS and NLLS estimates, which suggests the linear model places too much weight on large levels of R&D capital. Even with this alternative specification, the estimates can be taken as evidence that market valuations reflect a very positive impact from R&D capital of Spanish firms. Possibly, the main circumstance underlying these results is that R&D activities exhibit a particularly high correlation with size and the presence in the stock market, for the Spanish sample of firms compared with other countries. Also, estimates should be taken with caution because of the limited number of observations in our sample.

Investment and Market Value: Non-Quoted Firms

We have tried to explore in further detail the role of asset quantities for market valuation for firms publicly traded in the stock market and the rest of firms. Theory predicts intangible capital worths more for quoted firms because these assets are more important for equity holders, whereas tangible capital worths more for non-quoted firms because of debt holders. However, the number of traded firms in our sample is too tiny to figure out reliable results.

We alternatively explore the market value properties of the subsample that excludes traded firms. Therefore, we can compare the results for the whole sample in Table 8, with these additional estimates in Table 12, for the subsample of non-traded firms. The results in general do not change much. Actually, the valuation of firms in relation to their measured intangible capital increases. This finding is possibly revealing a strong correlation between the stock of measured intangibles and unmeasured and larger stocks of intangible capital. As stated in Hall (2001), it is not that the market values a euro of measured intangibles at 10 euros. Rather, the firm that has a euro of measured intangibles typically has another 9 of related unmeasured intangible assets. Finally, the fixed efect regression is not precisely estimated but for the efect of the additional other assets on the value of the firm, that apparently have even larger variability in annual realizations for this subsample.

[INSERT TABLE 12 ABOUT HERE]

Market Values on Asset Quantities: Non-Quoted

Summarizing, the results suggest that intangible investments have a positive impact on market values which is more substantial for innovative sectors, and this is also the case for R&D capital. Such a positive impact is influenced by the size of the firm and its presence in the stock market. An alternative explanation to low R&D intensity could be found in the small fraction of firms publicly traded in the stock market in Spain rather than in any small returns. It is commonly understood that new firms in innovative sectors with a expansionary potential have the stronger incentives in going publicly traded. This seems to be the case more in the US and UK than in continental Europe notwithstanding. Consequently, the evidence obtained from market values for innovative sectors in Spain suggests that promoting a more active role of market valuations as a guide for innovative investment might be a promising policy.⁶

4 Discussion and Concluding Remarks

A number of recent papers have examined the quantity, and the role of intangible investments for the US Economy. Part of these papers have focused on the market valuation of those investments. Here, we have addressed questions related to the measurement and valuation of the holdings of various types of capital for the Spanish economy. We have done this with an emphasis on innovative investments and notably on R&D capital.

In assessing the investment position of the Spanish economy in recent years the main finding is the relatively low level of R&D investment. This is not new. The striking result is that this figure seems to be relatively low compared to other intangible capital investments. Nevertheless, we find that the aggregate block of Scientific and Creative Property and Innovative Property is exhibiting the higher growth rates in recent years among the diferent categories of intangible investments.

It is worth noting that the aggregate measure of intangible investment we report for the Spanish economy is roughly consistent with the economic and technological gap we could more generally observe relative to the US economy. We can interpret this finding as if technological advances would require the same spending on intangibles irrespective of the degree of R&D intensity in the economy. This amounts to say that the technology adoption process, compared to the creation process, is relatively more important for the Spanish economy than it is for the US. The question is then whether the low R&D intensity is an indicator for poor firm’s expected economic results from innovative investments.

⁶See Farinós and Sanchís (2009) and the references therein on the limited stock market amplitude in Spain, as well as following Pagano et al. (1998) on the reasons why Spanish firms go public.

Our assessment for the relative worth of alternative investments uses the market value of the firms. The Central de Balances database which is available at the Bank of Spain is a very convenient source of data for this purpose since it pulls together firm’s market valuations with financial statements of tangible and measured intangible assets. We run regressions of market values on asset quantities over the period 1992-2005. The main finding is that the coeficient of the firm’s R&D capital is greater than one in our sample. If we assume that financial markets are eficient this finding suggests that firms disclosing R&D expenditure are investing too low, because the value of the assets created worth more than their cost. Note this result is for the business non-financial sector so that the average return of R&D investment could be reduced if publicly funded R&D yields lower return. The role for market valuation of institutions as, legal regimes, the stock market amplitude or ownership structures is an important issue that is left for further research.

It is worth noting as well, that the finding of a coeficient for an asset which is substantially above unity mostly reveals a strong correlation between the stock of that asset and unmeasured and much larger stocks of intangible assets. We have revised some of these candidate assets that can be proxied from various spending data from innovation surveys, and in sizeable amounts as documented along Section 2, for instance through the new organizational structure. Some other are relatively missing from our categorization like new market strategies and related. Finally, market values typically capitalize long-run characteristics of firms. Therefore, conditional on fixed efects, it is hard to find a significant role for annual realizations of tangible and intangible assets in afecting market values. This and other relevant econometric issues are left in part for subsequent research.

Interestingly, the patterns for the value of diferent investments are preserved if assets are interacted with sectors. Moreover, assets that can be associated to innovative investments have a stronger impact on market values in high and medium technology sectors. We think this finding supports the empirical strategy in this paper, and reinforces the interpretation on the Spanish economy being relatively more involved in technology adoption in recent years. Spanish firms might not be moving the technological frontier indeed, but it is very important to know how far they are from it. For this, further research is needed on the size, extent and difusion in Spain of technologies created by others. Existing and ongoing studies on R&D activity are crucial to learn on how much Spanish firms will be able to reduce such a distance with the technological frontier instead. Looking simultaneously to creation and adoption constitutes a hard but surely promising research agenda.

The results obtained in this study could be of interest for related applications and provide relevant evidence for the Spanish economy in light of the debate for the Lisbon agenda and the financing of innovative activities in the European Union.

References

1. Abowd, J.M., J. Haltiwanger, R. Jarmin, J. Lane, P. Lengermann, K. McCue, K. McKinney and K. Sandusky (2005), “The Relation among Human Capital, Productivity, and Market Value: Building Up from Micro Evidence.” In Measuring Capital in the New Economy, C. Corrado, J. Haltiwanger and D. Sichel, Editors, The University of Chicago Press.
2. B de E (2005), “Valoración de las acciones y otras participaciones en las Cuentas Financieras de la economía española.” Dto. de Estadística y Central de Balances, Banco de España.
3. Baily, Martin (1981), “Productivity and the Services of Capital and Labor.” Brookings Papers on Economic Activity: Macroeconomics 1:1981, 137-199.
4. Black, Sandra, and Lisa Lynch (2005), “Measuring Organizational Capital in the new Economy.” In Measuring Capital in the New Economy, C. Corrado, J. Haltiwanger and D. Sichel, Editors, The University of Chicago Press.
5. Brynjolfsson, Erik, Lorin M. Hitt and Shinkyu Yang, 2002, “Intangible Assets: How the Interaction of Computers and Organizational Structure Afects Stock Market Valuations.” Brookings Papers on Economic Activity: Macroeconomics 1:2002, 137-199.
6. Corrado, Carol, Charles Hulten and Daniel Sichel (2005), “Measuring Capital and Technology: An Expanded Framework.” In Measuring Capital in the New Economy, C. Corrado, J. Haltiwanger and D. Sichel, Editors, The University of Chicago Press.
7. Doraszelskiy, Ulrich, and Jordi Jaumandreu (2007), “R&D and productivity: Estimating production functions when productivity is endogenous,” mimeo.
8. Eaton, John and Samuel Kortum (1999), “International Technology Difusion: Theory and Measurement” International Economic Review 40(3): 537-570.
9. Estrada, Angel, Francisco de Castro, Ignacio Hernando y Javier Vallés (1997), “La Inversión en España.” Banco de España, Estudios Económicos, n 61.
10. Farinós, José E. and Vicente Sanchís (2009), “Factores Determinantes de la salida a Bolsa en España,” WP2009-03, IVIE.
11. Griliches, Zvi (1979), “Issues in assessing the contribution of R&D to productivity growth.” Bell Journal of Economics 10(1), 92-116.
12. Hall, Bronwyn (1993), “The Stock Market’s Valuation of R&D Investment during the 1980’s.” American Economic Review (Papers and Proceedings) 83(2): 259-64.
13. Hall, Bronwyn, and Rafaele Oriani (2006), “Does the Market Value R&D Investment by European Firms? Evidence from a Panel of Manufacturing Firms in France, Germany, and Italy.” International Journal of Industrial Organization, forthcoming.
14. Hall, Robert E. (2001), “The Stock Market and Capital Accumulation.” American Economic Review 91(5):1185-1202.
15. Jovanovic, Boyan (1997), “Learning and Growth,” in Advances in economics and econometrics: theory and applications, Vol. II,Kreps and Wallis eds. Econometric Society.
16. Lacuesta, Aitor, Teresa Molina and Luis A. Puch (2008), “Intangible Capital, the stock market and the Value of Spanish Firms,” mimeo.
17. Licandro, Omar, Reyes Maroto and Luis A. Puch (2005), “Innovation, Machine Replacement and Productivity: Evidence from Spanish Firms.” DP 5422, CEPR.
18. Licandro, Omar and Luis A. Puch (2008), “Measurement and Value of Innovation in Spain: A Contribution to the Energy Debate.” Focus-Abengoa Report, mimeo.
19. McGrattan, Ellen, and Edward C. Prescott (2005), “Taxes, Regulations, and the Value of U.S. and U.K. Corporations.” Review of Economic Studies 72, 76796.
20. McGrattan, Ellen, and Edward C. Prescott (2006), “Why Did U.S. Market Hours Boom in the 1990s?” Staf Report 369, Minneapolis Fed.
21. OECD (2008), “Intellectual Assets and Value Creation,” OECD Synthesis Report.
22. Parente, Stephen, and Edward C. Prescott (2000), Barriers to Riches. The MIT Press.
23. Romer, Paul (1990), “Endogenous Technological Change.” Journal of Political Economy 98(5): 71-102.
24. Pagano, Marco, Fabio Panetta and Luigi Zingales (1998), “Why do companies go public? An empirical analysis.” Journal of Finance 53: 27-64.
25. Solow, Robert (1957), “Technical Change and the Aggregate Production Function.” Review of Economics and Statistics 39, 312-320.
26. Solow, Robert (1960), “Investment and Technological Progress.” In Mathematical Methods in Social Sciences, K. Arrow et al. Editors, Stanford University Press, 89-104.

rce. CNE ^(IN Notes.(1)CNE,baseyear1995.Fromyear2000,CNE,baseyear2000.(p)Provisionalestimate.(a)Advanceestimate. Fixed Investm^{ent (Levels &} Ratios ov _{purchaser’s} price ₍m^illions of xed Capital Form^{ation (GFCF)}

rce. CNE ^(IN Notes.(1)CNE,baseyear1995.Fromyear2000,CNE,baseyear2000.(p)Provisionalestimate.(a)Advanceestimate. _mm^aterial Investment ^{(Levels &} Ratios ^o

ntific and Creative P^roperty, and Innovative Prope

ific and Creative Property, and Innovative Property ₍R^at

Economic Competencies ⁽

Economic Competencies ^(Ratios ov

ssions of M^arket Values on Asset Quantities, by sector of act