SECTION 2: INTRODUCTION

This section will focus on the characteristics of corporate research and development (R&D) and plant and equipment expenditures (PPE) categorised by firms financial constraints in the UK and if these financial constraints are really major determinants in firms investment activities. I will review neoclassical model of investment and base the assumption on imperfect capital markets rather than the usual perfect capital market assumption because of the feature of R&D as an intangible assets not having a collateral value and leading to problems such as moral hazard, adverse selection and asymmetric information which could have an impact on the limit of external finance. Also, looking at the empirical evidence on institutional shareholders to ascertain if they take a myopic view on long term projects or force managers to take a myopic view on long term projects. I have reviewed many different resources and journals on this topic and I will be comparing my findings with my empirical results on UK industries in subsequent paragraphs.

Leading British companies are reducing their R&D expenditure compared to their international competitors. According to recent publications in the Financial Times, February 3, 2010 they had been recent trends in the cutting of R&D spending in the UK, this has lead not only to cut down in investment but has lead firms to focus on projects that would yield a better outcome therefore leading them to ration its level of spending on long term projects that they think would yield positive returns. Below is a graphical representation showing the trend in R&D expenditure in the UK during 2000-2007,the horizontal axis measures the years and the vertical axis measures the mean values of R&D expenditures, there is a downward trend with the lowest bar showing the beginning of the financial crisis.

CHARACTERISTICS OF RESEARCH AND DEVELOPMENT AND PPE EXPENDITURES

I have studied R&D and also PPE expenditure, the first task is to look at the two different forms of investment and see if one is more subjected to financial constraints than the other. The second task is to compare existing literatures on physical and R&D investment under capital market imperfections and how they attract various types of financing.

Studying the behavioural pattern of R&D investment with different financial constraints is very useful because its bring new knowledge which affects long run economic growth, economic development and also long run success for fims.They are internal and external sources of finance for R&D and PPE expenditure .A firm that depends only on internal finance for R&D expenditure with little or no access to external finance , when faced with financial constraints or faced with an economic downturn that affects its cash flow will lead to a reduction in R&D expense and a fall in long run economic growth. Getting measures appropriately for financial constraints is not quite easy and has been a contentious issue but all through this paper, I have classified firms' constraints according to Ozkan (2002) who followed Fazzari et al (1988).

The most observable feature of R&D investment would be that it is classified as an intangible asset and therefore have no collateral value which makes it different from PPE which has a collateral value and is categorised as a tangible asset. It is also a very risky form of investment in the sense that there is high uncertainty in future cash flows and also in its success. It is difficult to know how much input would be needed for an R&D investment and what output to expect therefore it has a one-to-many relationship rather than a one to-one in a standard investment appraisal. With high uncertainty in success of R&D investment, it will be that for financially constrained firms with no access to external source of finance R&D investment could be highly sensitive to internal finance and a small change to internal finance either positive or negative would lead to a great change in R&D expense.

R&D investment requires a high level of specialization from individuals who are working on the project, and most of R&D spending is on wages and salaries of the specialised individuals. The firms' knowledge from experts creates the intangible asset (innovation), which would be their main source of future earnings if the project succeeds (innovative ideas of product or services or creation of new products and services). On the other hand, Knowledge is highly rooted in the firms employees and therefore will be lost if they are fired or leave. One of the implications of this feature is that it leads to high R&D adjustment costs because when firms' employees leave they take away essential part of the firm's resource base, so firms will have to smoothen their spending over time, in order not to lose more knowledgably workers. When firms are faced with a cut down in financial resources, R&D expenditure are likely to be cut down slowly, which points out the fact that firms smoothens R&D to some certain level when faced with temporary finance shocks.

Another feature is that it involves some level of secrecy between the firm and the potential investors who wants to finance the project or currently financing the project, reason is that innovative ideas would not leak to its rivals and competitors in the market place. This feature causes problem of asymmetry information between the firm and creditors or investors about the success of the project and leads to inaccurate appraisals on value of project by creditors. Despite all that, the feature (innovation) described above on R&D investment is non-rival in the sense that after a firm comes out with its innovative ideas, products and services, it could be imitated. Therefore, firms carrying out R&D investments would be discouraged to some certain level as its returns aren't fully enjoyed before another firm imitates the product, which will lead to a decline in R&D investment in the economy as many firms would not be interested in investing huge amounts into research and development projects. This had been articulated by Arrow but had been empirically discovered by Levin et al (1987) and Mansfield et al. They discovered through a survey, that it would be less costly to imitate a newly invented idea than discovering a new one. On the other hand, patents could be used to enjoy benefits from investment but had been viewed as ineffective by some firms in most industries according to Levin, Klevorick, Nelson and winter (1987) and would rather prefer secrecy. A current example where patents are not seen as effective is Pfizer cutting billion in annual R&D spending to 9.6 billion and would still continue this cut as a result of loss of patents, in the industry on blockbuster drugs and Pfizer Lipitor which is the best selling drug for the company is expected to face a generic competition late next year this was written in the Wall street journal February 2010.

THE ROLE OF CAPITAL MARKET IMPERFECTIONS AND EFFECTS ON HIGH TECH INDUSTRIES INTERNAL AND EXRTERNAL SOURCE OF FINANCE

This section looks into how capital market imperfections affect high tech industries. I have presumed that based on the data given that all firms are classified as high tech firms since R&D expenses are greater than zero. I choose to clarify the above so that my heading would not be misleading and the high tech sector provides an ideal platform in discussing the main idea on the literature of imperfections in the capital market

There are mainly three reasons why high tech industries with high R&D investment are mainly affected by imperfections in the capital market. Firstly, because of the nature of R&D investment having limited collateral value .Secondly, the gains from R&D investment are highly uncertain, skewed and tend to have low probability of potential future earnings or sucess.Lastly, there tend to be some informational asymmetries between firms and potential investors. Firms employees tends to have better information than outsiders or investors ,Cornell and Shapiro(1988, p.14) described the problem accordingly using this statement .The credibility gap between management and investors is likely to be most pronounced in the case of growth companies because management in such cases will often have better information about future profitability of undeveloped products and untapped niches .This great possibility for information increases the amount by which investors will discount the price of new corporate securities to compensate for information disadvantages.

THE ROLE OF DEBT FINANCING IN HIGH TECH INDUSTRIES AND ITS EFFECT ON R&D EPENDITURE

Additionally, focusing on firms access to debt financing goes along with the literature of capital market imperfections and according to Hubbard's (1998) empirical investigation of the literature ,it is seen that he's discussion of financial constraints is mainly drawn from the relationship between R&D's non-collateral value and the cost of debt. Debt financing increases the probability for financial distress especially in small high tech industries and the terms and conditions structured with it is not suitable for R&D investments due to uncertainty and unexpected returns. But we observed in section 1 that debt financing was insignificant with panel of small firms who had positive and negative cash flows in the UK industry suggesting the fact that small tech firms probably engage in the use of equity finance for R&D investment and this is consistent with Carpenter and Petersen (2002) who finds that equity financing has risen in countries with well- developed capital markets and also it could be due to the reasons cited above. In contrast, it's also observed based on our findings in section 1 that equity is insignificant in all our regressions also consistent with Carpenter and Petersen (2002) that found from their empirical investigation with US manufacturing companies that after firms go public they tend to depend on their internal finance for investment. The possible reason for the insignificant equity coefficient could be for example that UK's financial system makes it very expensive for firms to have access to external source, which leads to execution of R&D projects only by firms that are confident with using their internal source of finance available to them. This makes the cash flow coefficients to be more of an indicator, to tell if firms' carry out an R&D project rather than it telling us the level of R&D expenditure carried out which would be interpreted as financial constraints are important for UK firms Bond et al (1999).It should be noted as well that firms in the UK who carry out R&D project face fewer constraints.

Below in figure 1 according to carpenter and Petersen (2002) who followed the diagram in Bernanke et al(1998) and Hubbard (1998) is a supply of finance schedule, showing at various level of internal finance measured by the proxy of cash flow its corresponding marginal cost at each level. Flow of finance is measured on the horizontal axis and marginal cost of internal finance on the vertical axis. An upward sloping curve on the supply schedule depicts the supply of debt finance.

Firstly, the use of debt financing in high tech industries may not be the right financial instrument for R&D investment, because of moral hazard problems. This problem increases as firms becomes more leverage and are more intensified with the lack of collateral value of R&D investment. According to Arrow (1962 p.153) he points out that the problem of moral hazard is related to R&D investments because output can never be predicted perfectly from the inputs. Adding to this line of reasoning, Carpenter and Petersen (2002, p.f57). When borrowers' returns are highly uncertain, extensive use of debt may provide negative expected returns to lenders. Also, Stiliglitz and Weiss (1981) observed that as interest rates rises ,borrowers that aren't monitored by lenders tend to use loans to carry out riskier and higher return projects, which increases the chances of being bankrupt and offers no offsetting gains to debt holders if success is achieved.

Secondly, adverse selection problem also affects the debt market, which is mainly caused by information asymmetries because firms have better knowledge about the riskiness of the project than the lenders and this is more obvious with high tech investment than physical investment or any other type of investment. With such situation, lenders (banks) are more likely to ration credit rather than increase the interest rates so that low risk borrowers would also participate in the application of loans and to not exacerbate the situation of adverse selection. Also at low levels of leverage the debt supply schedule could be vertical as a result of credit rationing. On the other hand, debt financing is faced with the problem of moral hazard .High tech firms do have the ability to substitute between high risk projects and low risk projects and when creditors notice such behaviours, they ration credit or better still improvise some terms attached to the debt which definitely limits the firms behaviour. In conclusion it would be as moral hazard problems increases with firms leverage, restrictions or terms improvise by lenders on debt financing becomes more severe as leverage increases. A solution suggested by Berger and Udell (1988) is that small high tech firms financing could be resolved by mutual relationship lending with the borrower, they explain further that banks would get information over time through contacts with the firm, its local community which implies that small business tend to have better relationship with their borrowers but might not necessary be true for all sectors of the economy.

Thirdly, one of the main features of R&D investment as earlier stated is its no collateral value and therefore limits access to debt financing. Based on empirical investigation by Bester (1985, 1987) it could be citied that collateral is a form of signalling to lenders differentiating a low risk project from a high risk borrower and is a device to reduce the problems of moral hazard. Berger and Udell (1990) says “Collateral plays an important role in the U.S domestic bank lending ,as evidenced by the fact that nearly 70% of all commercial and industrial loan are currently on secured basis”. Due to the fact stated above, that is why equity and not debt financing is usually citied as the best instrument for financing R&D projects.

Lastly, the expected marginal cost tends to increase with high use of borrowed money for R&D investments especially in high tech industries .In section 1 we also observed the debt coefficient significant for large firms and insignificant for small firms, so this shows the fact that small high tech firms in the UK do not depend on debt financing for R&D investment as this increases their risk of bankruptcy. Also we could interpret the significance of debt financing in large firms as these firms acquiring physical properties over time which had been used as collateral to secure their loans from lenders. Financial distress could lead to the loss of vital employees and abandoning critical projects. Brown (1997) argues on the fact that tests, which are used to detect the impact of capital market imperfections in high tech firms are not able to distinguish between the two possibilities of either capital markets being perfect or imperfect and the various factors, which encourages different types of investment to react differently to a common factor which could be shocks to internal finance. He then carries out a test to compare sensitivity of R&D investment to cash flow and he came up with the result that capital markets are imperfect and that investment for high tech firms is highly sensitive to cash flow.

ROLE OF EXTERNAL FINANCE ON R&D INVESTMENT

Equity financing could be seen as a solution to debt finance for R&D investment especially when firms are young (carpenter and Petersen (1998)), who concluded in his empirical investigation of 2400 US high tech firms that the growth of young high tech firms were observed after their initial public offerings and also as the firms began to grow most of its expenditure is being financed with its retained earnings. Equity financing either internally or externally does not require any form of collateral and does not lead to financial distress of any sort, and does not create problems of adverse selection. Equity finance is very expensive, given the amount of shares needed, the premium on equity varies inversely with the borrower's sum of internal finance and the collateral value of liquid assets and would we therefore observe that equity finance would be expensive for financially constrained (Ozkan, 2002).On the other hand, the most cheapest source of finance for any firm would be cash flow as proxy for internal finance and if there's no availability either due to financial constraints firms turn to debt financing as an option, which rarely available or finally issuing new shares. In conclusion, the fact as established in aforementioned paragraphs do suggest that as debt financing becomes difficult to access, and the only way for firms to achieve rapid growth and enlarge its firm size would only be possible if firms have access to equity markets. Factors which affects R&D spending in the UK is the level of cash flow or internal finance available ,access to credit (debt financing),and good financial markets. If all these factors are intact we would expect a positive increase in R&D expenditure but with the recent economic downturn access and availability of sources of finance had been reduced.

THE ROLE OF INSTITUTIONAL OWNERSHIP ON LONG TERM PROJECTS

Institutional owners are seen to have short sightedness towards firms engaging in R&D projects ,therefore only interested in short term projects which are value enhancing in the short run but that's is not always the case as others have found that institutional investors to increase the level of R&D investment in companies they have a stake in. According to Wahal and Mc Connell (2000), the functions of institutional owners is that they act as a buffer between raring to go individual shareholders and corporate managers and, thereby allowing corporate managers to focus on projects with long term payoff, basis for such a view is that institutional investors have better information advantage compare to share holders. Without doubt, institutional owners are rarely seen to evaluate corporate manager's based on reports of short term earnings compared to individual shareholders. In section 3 I would go into more details to see if institutional ownership has a positive effect on R&D expenditure which Wahal and Mc Connell and also Jarrell et al discovered in their papers.

SECTION 3. INTRODUCTION

This section would be a further examination on the empirical analysis between R&D expenditure and institutional ownership in the UK to see if there's positive relationship incorporating that institutional ownership encourages R&D investment in the UK, similarly for PPE .On the other hand, I would classify firms further into financially constrained and unconstrained and investigate on how determinants of R&D and physical investment differ in this two groups. Estimation techniques used would be the fixed effects estimation method, which would be applied to the regression done in section 1 and also to the firms' classification of financial constraints. Moreover, I classify firms as finically constrained in the light of Ozkan (2002) and Fazzari et al (1998), whereby using total leverage ratio (total debt-to-total assets) and if above its median classified as financially constrained and when the total leverage ratio is below the median or the 50th percentile value of 0.1061549 classified as financially unconstrained. I choose to use this criteria in classifying because this indicates the likelihood of a firms financial distress, a high ratio implies high leverage and vice versa. This form of indication provides signal of creditworthiness could it be in the case of financial distress firms with low debt would have less problems in paying back its debts. On the other hand, most debt payments are usually with interest rates and higher debts means higher interest to pay and therefore existing interest commitments can influence firm's access to credit and therefore limit R&D investment (Sinai and Ecksten, 1983).Finally firms that do have high total debt ratio tend to have more sensitivity of R&D expenditure to cash flow, as chances of finding other sources of finance are minimised.

DATA DESCRIPTION

The data for this study are taken from World scope and Thomson One Banker. Our sample consists of 570 firms and 2788 observations covering 15 industries and a seven year period from 2000-2007.The methodological approach followed is according to Wahal and Mc Connell (2000),the econometric specification is given below, where all variables had been scaled down by total assets expect for institutional ownership,QIt. The transformation of the model makes it become ratios rather than levels because this helps us to compare investment and R&D expenditure ratios over time and across firms. In a panel with firms growing over time as well as different sizes such as a transformation yields trend stationary series and controls for heterogeneity (Himmelberg and Petersen 1994).The net debt coefficient which is a flow variable would be used in the fixed effect regressions as these gives current value of what the firms owes rather than accumulated debt.

R&Dit = aindustry + ayear + b1CFit + b2Qit + b3Debtit + b4Equityit + b5InstOwnerit + b6 (InstOwnerit * CFit) + eit.

DESCRPTIVE STATISTICS

The descriptive statistics is both in tables 1 and 2,table 1 consists of our key independent and dependent variables used in our analysis under financial constraints and also all the sample without constraints. The mean of R&D expenditure scaled by total assets is significantly larger in financially unconstrained firms with mean of 0.135 and in financially constrained with 0.072 these would be as accepted as firms that financially unconstrained tend to have various option to sources of finance and therefore would invest more in R&D expenditure.

On the other hand, PPE expenditure is greater than R&D expenditure in all samples with a difference of 0.096 as would be expected because of the collateral nature of PPE so creditors would be more willing to give credit to fund long term projects such as PPE.The mean for PPE for financially constrained firm 0.267 is greater than that for financially unconstrained 0.135 a possible explanation for this could be reflected through their mean values of -0.835 for financially constrained and -0.149 for financially unconstrained which shows that financially unconstrained firms prefer to use most of its cash flow to finance PPE rather than R&D that has less chance of succeeding. The total debt for the classes of firm is as we expected mean of 0.363 for financially constrained and 0.024 for financially unconstrained which reflects the true position of the firms leverage. Lastly, the mean for the share of institutional ownership is 43.24 for financially constrained and 47.06 financially unconstrained this could imply that institutional shareholders prefer to invest in firms that have lower leverage. The table is reported in the appendix.

Table 2 shows the year-by-year means and medians of relevant dependent and independent variable .According to the first variable which is R&D expenses across firms per year, the distribution is highly skewed. In each year it's observed that the mean value of R&D expenses is 20 times the median. Data could be seen to have some times variation, but no pronounced times series trend (Wahal and McConnell) .When R&D expenses are scaled by total assets the means and median tend to be more right ward skewed with the highest mean value 0.130 observed in 2003 and a gradual decline of 0.99 in 2007 which is associated with the beginning of the financial crisis where most firms especially pharmaceuticals industries started cutting R&D expenses.

Also PPE shows similar trend as R&D expenditure, the distribution of the mean is highly skewed but once scaled down by total assets it becomes rightly skewed the mean of PPE goes down in 2006(0.163) but increase slightly in 2007 with 0.10.The mean of the share of institutional ownership is highest in 2006 with 56.15% but decrease with about 4% which again could be associated with the financial crsis.The findings of expanding role of institutional ownership is consistent with Davis(2002) who analysed the relationship between institutional ownership and corporate finance sector for OECD countries

VARIABLES

STAISTICS

ALL SAMPLE

HIGH TOTAL DEBT /TA RATIO

(FINACIALLY CONSTRAINED)

LOW TOTAL DEBT /TA RATIO

(FINANCIALLY UNCONSTRAINED)

R&D/TA

MEAN&MEDIAN

0.104

0.038

0.072

0.021

0.135

0.069

PPE/TA

MEAN

MEDIAN

0.200

0.148

0.267

0.221

0.136

0.083

CF/TA

MEAN

MEDIAN

-0.116

0.058

-0.835

0.672

-0.149

0.020

TOTAL DEBT/TA

MEAN

MEDIAN

0.194

0.106

0.363

0.260

0.024

0.004

EQUITY/TA

MEAN

MEDIAN

0.116

0.001

0.069

0.001

0.162

0.003

INST.OWN.

MEAN

MEDIAN

43.24

42.5

47.06

46.28

39.75

38.54

TABLE1 SHOWING DESCRIPTIVE STATISTICS OF FINANCIAL CONSTRAINTS

TABLE 2 (SHOWING VARIABLES YEARLY MEAN AND MEDIAN PATTERN

STATIS.

2000

2001

2002

2003

2004

2005

2006

2007

R&D exp.

MEAN

MEDIAN

40.100

2.348

35.259

2.056

33.471

1.746

30.231

1.417

28.155

1.38

26.090

1.229

26.067

1.112

23.751

0.913

R&D/TA

MEAN

MEDIAN

0.067

0.023

0.079

0.031

0.094

0.041

0.130

0.049

0.123

0.047

0.105

0.042

0.108

0.034

0.099

0.032

INST.OWN

MEAN

MEDIAN

36.469

36.67

37.220

36.505

37.679

38.98

38.775

39.03

38.878

41.32

41.351

43.01

56.515

54.945

52.114

51.6

PPE

MEAN

MEDIAN

869.19

19.500

826.90

9.603

814.32

6.687

726.82

4.561

672.51

4.239

654.47

3.582

606.99

2.483

669.29

2.800

PPE/TA

MEAN

MEDIAN

0.253

0.219

0.229

0.182

0.216

0.175

0.214

0.169

0.204

0.148

0.181

0.125

0.163

0.110

0.173

0.100

METHODLOGICAL APPROACH (DESCRIBING THE FIXED EFFECTS AND RANDOM EFFECTS MODEL)

Our Sample as mentioned above is a time series of cross sections using this form of sample have its advantages and disadvantages .One of the main advantage of using this sample is that the time series element in the sample gives us the opportunity to model dynamic behaviours and also cross sections allows for the behaviour of different individuals to be modelled and reduce problems of aggregations, the disadvantage of using panel data is that it causes problems of hetreoscedasticity, simultaneity and serial correlation and also attrition problems. , I have used robust standard errors to take care of the problem of heteroscedaticity in the pooled OLS regression and fixed effects regression as well. I have used both the within and between fixed effects estimation techniques to run my regression.

Additionally, panel data samples deals with unobserved heterogeneity as in the case of our sample, unobserved heterogeneity is the firms specific fixed effects which could include different culture of companies and these firm specific effects stay constant over time .On the other hand, controlling for these firms specific effects are important as there could be a positive correlation between sources of finance and R&D. The mainly observable sources of the correlation could be that for firms they are different patterns in managerial style and good managerial style could bring about increased profits and cash flows and firms would tend to grow faster compared to bad managerial style and also some firms have better relationship with creditors therefore having more access to debt financing for investment in PPE. Failure to account for these could cause the estimates to be biased and it could be seen as a specification error. The specification of our model using the fixed effect method of estimation is specified below.

R&Dit = ayear + b0it + b1CFit + b2Qit + b3Debtit + b4Equityit + b5InstOwnerit + b6 (InstOwnerit * CFit) + eit .... (i)

Where b0 is the individual firm's specific effect and ayear is the year effect. The method used to clear the firms fixed effects is transforming the variables to deviations from their specific means, error term accommodate the measurement error in the dependent variable and the effect of unobserved dependent variable assumed to be uncorrelated with cash flow and firms and year effects. Below shows the equation when observations of each cross sectional unit are summed over the time dimension and divided by T.Type equation here.

it = ayear +0it+ b1it + b2it+ b3it + b4it + b5it + b6() + it..........(ii)

Deducting ( ii) from (i)

R&Dit - it = b0it -0it + b1it)+ b2 ( Qit -it) + b3 (Debtit -it )+ b4 (Equityit - it )+ b5 (InstOwnerit - it )+ b6 (InstOwnerit * CFit- ) +( eit -it).

could be estimated by OLS and gives the within estimator or fixed effects estimator controlling for firms specific effect using the command of xtreg, FE robust command correcting also for hetreoscedaticity.The demerit of using this estimator is that if any of the variable do not vary over time but only in cross sections, the variable would be perfectly collinear with unobservable firms fixed effect coefficients and the coefficients would therefore not be estimated. I have also used the between firm estimator despite it doesn't control firm effects and also less efficient, the reason behind using these estimation technique is that according to Himmelberg and Petersen (1994) “is that the transitory component of cash flow tends to average out over time and hence provides the extent to which the within firm estimates are biased downwards”. The approach to obtaining these estimates is to regress the mean of the fixed effects dependent variable on the mean of the specific fixed effects of the explanatory variable. If there's a difference between the within and between estimates we would regard that they might be misspecification in the model.

In conclusion, the FE poses a problem of correlation between fixed effects and error term an alternative approach could be using the random effects buts these estimation technique has problems of endogeneity as well therefore causing estimates to inconsistent in large samples. The RE could be calculated as follows where

Assume

RE results are weighted average of between the within group (FE) estimator variance of. Between group estimators, where it has a distinct intercept and it is an estimator based on the variation between group mean variance The hausman test would be used to test for endogeneity of the random effects, where the null is that there's no endogeneity and under HO the RE estimator is consistent and asymptotically efficient and the FE is consistent, under H1 the FE estimator is consistent and the RE estimator is inconsistent. Hausman Wald test = where =

EMPIRICAL RESULTS COMPARISON OF SECTION 1 POOLED OLS WITH FE

Comparing the results of section 1 using pooled OLS and fixed effects(within) estimation technique shows differences in the estimated coefficients(table would be included in the appendix). A reminder that the classes of firm based on financial distress had been classified with using total assets greater than median as large and vice versa for classification of small firms. The cash flow coefficients for sample of all firms using net debt are respectively (-0.035 ) using fixed effects and (-0.037) with OLS the fact that the coefficient of the FE estimate has reduced compared to the OLS results suggests that the latter may be biased upwards and for all estimated coefficients of cash flow in both large(all sample CF>0 and CF <0) and small firms(all sample except for CF>0 & CF<0 as fixed effects has higher coefficients, CF>0 coefficients with FE is (0.143) and (0.097) and CF<0 with FE (0.045) and OLS (-0.021) respectively) .We observe a statistically significant coefficient and positive at 5% level of significance for cash flow using FE for large firms that have CF>0 (0.113) therefore having a positive effect on R&D which agrees with Ozkan (2002) and Himmelberg and Petersen (1994)who find similar positive relationships. Contractdicting FE is significant, but have a negative coefficient in all sample for small firms (-0.043) and lastly in panel of all firms where total debt and net debt is a proxy for debt financing (-0.046) therefore not supporting R&D expenditure.

Lastly for the comparison of section 1 pooled OLS result with the FE, institutional ownership is both significant and negatively signed using OLS and FE with panel of all samples where total debt is a proxy of debt financing also with panel of all small firms with coefficients of (-0.001) respectively, interpreting that institutional ownership is significant to determine R&D expenditure but have a minimal negative impact towards R&D as coefficients are quite small.R2 (1-RSS/TSS) tends to bigger in Pooled OLS regression this is because we used industry and year dummies in running the regression , but we used only the year dummies in the FE regressions so they are more explanatory variables in the pooled OLS therefore reducing RSS and increasing R2 .All coefficients are jointly significant.

In summary our results with FE shows that significant determinants of R&D are institutional ownership, cash flow, Q, equity. There's no significant coefficient of total debt using FE but only in OLS where its negatively significant in large firms and classification of CF<0 in large firms ,which tells us that after controlling for firms specific effects debt financing is not a method of financing R&D expenditure and this is consistent with the feature of R&D not having a collateral value and also its low probability of success and therefore debt financing is not a source of finance which is consistent with the findings of Carpenter and Petersen(2002).In OLS regressions we discovered that cash flow is not significant in all panels which was consistent with Bond et Al(1999) who found that cash flow was not important for R&D spending for UK companies but tells us if the firm takes up R&D investment and not the level of investment. In conclusion I would draw up that results that we get from pooled OLS does not take into consideration different firms specific effects which could be correlated with the various sources of finance, and the FE takes these into account so I would rely more on the results of FE as our panel is looking at different industries and all the unobserved effects tends to affect estimates. In the next subsection I would look at the difference between within and between estimates of the fixed effects estimation technique.

TABLE 4: REPORTING FIXED EFFECTS(WITHIN AND BETWEN) REGRESSION WITH R&D AND PPE

EXPLANATORY VARIABL-E

ALL FIRMS

WITHIN

ALL FIRMS

BETWEE-N

R&D

F.C

R&D

F.UNC

R&D

ALL FIRMS

WITHIN

PPE

ALL FIRMS BETWEEN PPE

F.C

PPE

F.UNC

PPE

CFit

-0.035

(-1.81)

-0.021

(-1.31)

-0.024

(-1.42)

-0.043

(-1.48)

-0.00009

(0.03)

-0.044

(-194)

0.001

(0.05)

0.0021

(0.24)

Qit

0.005

(3.52)*

0.013

(8.58)*

0.003

(0.60)

0.0038

(2.00)*

0.001

(2.51)*

-0.006

(-2.84)

-0.001

(-0.58)

0.001

(2.49)*

N.DEBTit

0.045

(0.77)

0.157

(2.25)*

0.024

(0.62)

0.004

(0.08)

0.063

(1.67)

-0.002

(-0.02)

-0.025

(-1.61)

0.028

(-1.19)

EQUITYit

-0.052

(-2.94)*

0.108

(3.93)*

-0.022

(-0.58)

-0.059

(-3.01)*

-0.023

(-3.10)*

-0.067

(1.70)

-0.0122

(-O.79)

-0.025

(3.80)*

INST.OWNit

-0.0006

(-2.91)*

-0.0003

(1.47)

-0.0002

(-1.47)

-0.0006

(-2.33)

-0.0001

(2.09)*

0.0007

(2.45)

0.0001

(-1.63)

0.002

(1.88)

Inst*cf

-0.0014

(-2.67)*

-0.0025

(-5.61)*

-0.0005

(-1.43)

-0.0020

(-2.19)*

-0.000

(-2.66)*

0.002

(3.35)

-0.0003

(-1.93)

-0.000

(-1.20)

No of obs

2400

2400

1142

1258

2398

2398

1141

1257

R2

W 0.257

B 0.448

0.1589

0.3257

W 0.1197

B 0.1146

0.1540

0.0758

F-STATISTIC

9.07**

31.43**

1.86**

6.30**

11.32**

6.53**

8.84**

4.05**

Note: Estimated with year dummies (not reported).

Heteroskedasticity consistent t-statistics reported in brackets.

*Variables are significant at 5% level.

**F-statistics critical value of 1.5

FC-Stands for financially constrained measured as total debt greater than median

FUC-Stands for financially unconstrained measured as total debt less than the median

Under R2 W stands for within R2 and B stands for between R2

WITHIN FIRM AND BETWEEN FIRM RESULTS

Using the within estimates takes care of problem of unobserved heterogeneity, as there might be a possibility of firms specific effect being correlated with sources of finance and R&D. Firstly we need to test for endogenity between fixed effects and error term using the hausman test procedure which had been explained above the test results shows when R&D is the dependent variable for panel of all firms using the within chi2 (13) =312.92 Prob>chi2 = 0.0000 panel of all firms using the between chi 2(12)=151.09 prob> chi2 =0.000 so we reject the null so FE is consistent.

There is a difference observed between the within estimates and the between estimate in the R&D regressions in panel of all firms and also PPE with the within estimates upwardly bias, this could be as a result of misspecification of the model .Another reason is that we have an unbalanced panel not all firms in various industries examined have enough information. Between estimates are less efficient and there tend to have larger variance which deflates t-ratios, but in this case we observe higher t-ratios in comparison to the within t-ratios .R2 tends to be larger in the between estimates regression than the within estimates regressions this is because the between regressions doesn't take deviations of the fixed effects but rather includes the mean of the fixed effects in the regressions.

The within and between estimate in regressions, where the dependent variable is R&D and PPE an insignificant coefficient is observed for cash flow suggesting that cash flow does not support R&D spending ,which is consistent with Bond et al (1999) who found similar results for UK firms. Additionally, I would not have expected such insignificant results for cash flow because according to financial hierarchy it groups cash flow as the cheapest source of finance for R&D.

Also, for institutional ownership I find the within estimate to be negatively significant with panel for all samples and panel when classified as financially unconstrained with coefficients of -0.0006 respectively when R&D is used as dependent variable . Which interprets as institutional shareholders not supporting R&D investment in UK companies due to the negative coefficient, this result is also consistent with Davis (2002), who finds that institutions appear to accompany lower investment in Anglo-Saxon economies. I find similar results for PPE with coefficient negatively significant for within estimate -0.001 in panel of all firms, but positive with between estimates in panel of all firms with coefficient 0.0007.

Debt coefficient is only positively significant for all panel samples with the between estimate in R&D regressions 0.045 and insignificant in all other samples. This result tells us that net debt as a proxy for debt financing is not used as a source of Finance for R&D in UK companies which is also consistent with Hall (1992) that found that debt is not used to finance R&D in US manufacturing companies .I also find net debt insignificant using PPE as dependent variable, in all regressions which is a little bit strange as debt financing is seen as a source finance to support PPE investment, this result basically shows that our data is highly skewed.

Equity is significant for R&D regression for panel of all firms with within and between estimates, negative coefficient with within estimate (-0.052) and panel of financially unconstrained with coefficient of (-0.059) but positive with between estimate (0.108). The insignificance I observed for constrained panel shows the fact that equity as a source of finance for R&D is very expensive. On the other hand a statistically significant negative coefficient is observed, when PPE is dependent variable in all panel samples with within estimate and with financially unconstrained with coefficients of -0.23 and -0.025 respectively which I would expect as PPE and equity are negatively correlated. Tobin's q is statistically significant in all firms panel when R&D is dependent variable with positive coefficient of 0.005 and 0.013 for both within and between estimates respectively. It is positive for PPE panel of all firms with within estimates of 0.001 and financially unconstrained with coefficient of 0.001 respectively.

LOOKING AT SELECTED INDUSTRIES

I decided to look at four industries in our panel of samples that are I believe needs continual R&D investment the industries are automotive, aviation and transportation which is classified as IND1,IND4 chemicals ,healthcare and pharmaceuticals, IND7 engineering ,mining and oil gas exploration, and IND8 food producer, processing and farming and fishing. Below is a graphical representation showing the pattern of yearly means of R&D expenditure. As expected R&D investment is needed for IND1 and shows an upward trend reaching its peak in 2006 and having a sharp decline in 2007, the next industry that shows high R&D expenditure is IND4 as expected followed by IND7and IND8.Looking further at FE regressions of this aforementioned industries, using R&D as dependent variable I found institutional ownership statistically insignificant in panel of all firms without any sort of financial constraints in IND4,7,8 with t-ratios of (-1.22,1.81,0.70) respectively, but found institutional ownership in IND1 negatively significant with coefficient of (-.00024) .When grouped into constraints I found only IND1 when grouped as financially constrained to be positively statistically significant with coefficient 0.0004 .Under financially unconstrained all coefficients are insignificant and IND8 having no information. Despite looking at industries we could conclude that institutional ownership in UK companies do not support R&D investment and are only concerned with short term projects which is consistent with Davis(2002) and wherever they are significant they have very small coefficients which may just cause little effects.

ALTERNATIVE ESTIMATION TECHNIQUE

An alternative method that could be used to obtain consistent estimates of coefficients in our model would be to follow the research strategy of Grilliches and Hausman (1986 p.114) whom Himmelberg and Petersen (2002) followed. The process would be to remove firm effects by first differencing and also dependent variable and endogenous explanatory variable (cash flow, debt, inst ownership) where the year dummies would be suppressed and consistent estimates of variables are obtained using IV.The instruments that could be used would be the lags of the endogenous explanatory variables up to t-2 ,which would be highly correlated with the first difference of current explanatory variables, but uncorrelated when a composite error term is used and is independently distributed. In the presence of serial correlation in cash flow, we include an alternative model that allows variables to follow an MA (1) process. This process could be estimated by GMM and we would have estimates which are consistent and robust to the choice of instruments. For the level of course GMM had not been taught.

CONCLUSION

I would conclude that institutional ownership and therefore could be seen as myopic towards the decisions of long term projects either PPE or R&D in the UK, if I was able to do the alternative estimation method mentioned above am sure I would have found more interesting results, most papers I read did the GMM method and found positive results for instititutional ownership in the US not having any form of myopia but encourages long term investment. Lastly it should be noted that our panel of sample is unbalanced as some information tend to be omitted from the data.

Source: Essay UK - http://turkiyegoz.com/free-essays/finance/characteristics-of-corporate-research-and-development.php


Not what you're looking for?

Search:


About this resource

This Finance essay was submitted to us by a student in order to help you with your studies.


Rating:

Rating  
No ratings yet!

  • Order a custom essay
  • Download this page
  • Print this page
  • Search again

Word count:

This page has approximately words.


Share:


Cite:

If you use part of this page in your own work, you need to provide a citation, as follows:

Essay UK, Characteristics of corporate research and development . Available from: <http://turkiyegoz.com/free-essays/finance/characteristics-of-corporate-research-and-development.php> [17-12-18].


More information:

If you are the original author of this content and no longer wish to have it published on our website then please click on the link below to request removal:


Essay and dissertation help

badges

 
Look Before You Leap Essay | LEGO Dimensions | Télécharger !