Organization Capital

Intangible assets—sources of future benefits that lack physical embodiment—are the hallmark of modern economies and business enterprises. Various estimates indicate the significance of such assets in the modern economy. For example, Nakamura (2000) estimates the value of U.S. corporate investment in intangibles during 2000 to be around $1.0 trillion, making it roughly equal to the total investment of the non-financial sector in property, plant and equipment. Hall (2000) estimates the total value of intangible capital as ranging between half to two-thirds of the total market value of publicly traded corporations, as indicated by the q ratio (market value to replacement cost of physical assets). Nakamura (1999, 2000) argues that the major growth in value and impact of intangible capital started roughly in the mid-80s, with the emergence of major ‘intangible industries’ (software, biotech, internet, etc.) and continues unabated, though with considerable fluctuations, to the present. Gu and Lev (2001) show that firm-specific estimates of intangible capital improve significantly the association between capital market values and accounting-based measures of performance and value (e.g., earnings or book values). More recently, McGrattan and Prescott (2007) emphasize the importance of considering intangible investments in explaining the real economic growth in the 1990s.¹ Overall, it is widely accepted that intangible assets are the major drivers of national as well as corporate value and growth, as most forms of physical and financial assets are essentially commodities, yielding on average the cost of capital.

A framework developed by Lev (2001) for intangible capital classifies intangible assets into the following four groups.

While economic and business research as well as popular managerial writings on the first three classes of intangibles (discovery, customers and human resources) are voluminous, systematic research and knowledge about organization capital is in its infancy.

It is widely observed that within industries some companies systematically outperform their competitors and maintain their leading position for long periods of time, despite significant economic changes: Wal-Mart in retail, Microsoft in software, Southwest among airlines, DuPont in chemicals, Exxon in oil and gas, Intel in microprocessors, and the list goes on. Such superior performance in terms of growth in sales, earnings and stock returns cannot be attributed to monopoly power or government subsidies because these firms operate in a competitive environment. How can such enterprises achieve/maintain their superior performance and leading role? We argue that such enterprises have a stealth asset: organization capital—the agglomeration of business processes and systems, as well as a unique corporate culture, that enables them to convert factors of production into output more efficiently than competitors. Anecdotal examples of such business processes and systems abound: Wal-Mart's supply chain, where the reading of barcodes of purchased products at the checkout register is directly transmitted to suppliers who are in turn responsible for inventory management and product provision to Wal-Mart stores; Cisco's internet-based product installation and maintenance system, estimated by Cisco's CFO to have saved $1.5 billion over three years (Economist, 26 June 1999); Zara, a successful clothes and accessories retailer, which transmits in real time customers' choices to its suppliers worldwide. This agglomeration of business processes and systems that cannot be easily mimicked by competitors is the intangible asset, organization capital.

Even though terms like ‘firm's reputation’, ‘value of leadership’, ‘capacity to innovate’, etc. that capture certain elements of organization capital abound in the management literature, little in terms of operating measures of organization capital and empirical evidence is available. In essence, there is sparse systematic evidence on the magnitude and impact of organization capital, in contrast to the substantial evidence on other intangibles, such as R&D, patents or brands.²Lev and Radhakrishnan (2005) develop a methodology to estimate organization capital and provide preliminary evidence on the role of this asset in market values. This measure is based on the contribution of organization capital to revenue growth.

In this paper, we extend the methodology developed by Lev and Radhakrishnan (2005) to estimate firm-specific organization capital, by incorporating a firms' potential to save operating costs as well. Thus, the new measure of organization capital incorporates a firm's potential to generate super-normal revenues as well as cost savings. We validate the measure by examining its association with future firm performance measures, such as operating income growth, sales growth and abnormal returns. We find that the organization capital measure is associated with five years of future operating and stock return performance, after controlling for other factors. Thus, the organization capital measure indeed captures the notion of a firm's ability to maintain its leadership position.

We also examine the association of the organization capital with executive compensation. Successful organization capital is in fact a reflection of managerial ability, since managers are responsible for the creation and maintenance of this asset. Executive compensation is an incentive mechanism as well as a measure of the executive's ability that should manifest in business processes and systems, that is, the way of doing business. We find that executive compensation is positively associated with our measure of organization capital. This is important because it suggests that organization capital proxies for the elusive concept of ‘managerial quality’. Thus, the compensation committee of the board can use the organization capital measure as an aggregate proxy for managerial ability and quality for compensation purposes.

BACKGROUND: WHAT IS ORGANIZATION CAPITAL?

Evenson and Westphal (1995, p. 2337) provide a succinct definition of organization capital: ‘organization capital . . . [is] the knowledge used to combine human skills and physical capital into systems for producing and delivering want-satisfying products’. Accordingly, organization capital relates to firms' underlying systems and processes of operating, investment and innovation capabilities that enable them to generate outputs. While all enterprises require a certain organization capital, to convert resources into outputs we use the term ‘organization capital’ to refer to the ability of firms to deliver and sustain super-normal performance. That is, organization capital enables superior operating, investment and innovation performance, represented by the agglomeration of technologies—business practices, processes and designs.

An important feature of organization capital as an enabler of super-normal performance is its tacit nature. In essence, for organization capital to provide firms with a competitive edge, the capabilities encompassed in business processes cannot be completely codified and transferred to other organizations or imitated by them. Evenson and Westphal (1995, p. 2213) emphasize this feature: ‘Much of the knowledge about how to perform elementary processes and about how to combine them in efficient systems is tacit, not physically embodied and neither codified nor readily transferable’. The retail industry demonstrates the difficulties in imitating and adopting others' organization capital. With all that has been written about Wal-Mart's business processes, such as vendor-managed inventory and supply chains, electronic data exchange systems, K-Mart was largely unsuccessful in imitating these business processes. Wal-Mart continues to be the leader in efficient supply chain management, as is Apple in innovation and Exxon in exploration. Overall, the knowledge or information underlying the business operations is not easily transferable across companies, even over extended time periods. This feature of organization capital is referred to in the management literature as management culture and capability and thus is recognized as an important intangible asset.

The tacit and idiosyncratic nature of organization capital makes it particularly hard to measure. Different enterprises can have different areas of superior capabilities. For instance, some enterprises have superior business processes and culture enhancing the innovation capabilities, while others may have superior processes to tap into the human capabilities. Loosely speaking, there are different roads that lead to a sustained superior performance, and a direct measure of organization capital is difficult to obtain.

Adding to this fundamental difficulty of measuring organization capital—its tacit nature—is the fact that the investment (input) in organization capital is not fully tracked by firms (e.g., the cost of on-the-job training) nor reported publicly. Even in the realm of innovation expenditures, such as research and development, firms at most provide the total expenditures, but no information is available on the nature of research and development activities. Therefore, a direct measure of organization capital based on inputs is not feasible. As such, our measure of organization capital provides a relative measure of firms' capabilities.

We proceed to develop the measure of organization capital.

FIRM-SPECIFIC ESTIMATE OF ORGANIZATION CAPITAL

All companies use similar resources (inputs) to generate revenues: physical capital, labour and periodic expenditures for generating and maintaining organization capital. While all companies use similar resources, there are significant differences across companies in the efficiency of use, or the contribution of the resources to revenues. For example, while companies A and B use employees, Company A's employees may generate more revenue than B's because they are better trained and compensated. Similarly, while both A and B have physical capital, Company A may generate more revenue per unit of physical capital than B because A uses more sophisticated IT and optimization models. In short, there are many reasons why companies differ in the efficiency of resource usage, but most of these reasons (better IT, higher quality employees, improved incentive and compensation systems) are related to the organization capital. Accordingly, the way to measure organization capital is to compare across companies the efficiency of using the resources in generating revenues as well as in cost containment. The procedure is described below.

Estimating Organization Capital

We model the firm's output—sales (SALE)—as a function of physical capital (PPE: property, plant, and equipment) and labour (EMP). We use for estimation the following constant returns to scale production function (see Hall, 2000):

(1)

where SALE_it is the revenues of firm i in year t, PPE_it is net plant, property and equipment, EMP_it is number employees, and e_it is an error term. The constant, a_0it is the productivity parameter, which we model as a function of the instrumental variable, SGA_it as follows:

(2)

where SGA_it is the sales, general, and administrative capital, computed by capitalizing and amortizing the annual SG&A expense over three years. The firm's SG&A expenses include outlays related to the building of organization capital, such as information systems, employee training, research and development, consultants' fees, and brand promotion. We allow for two types of contributions of organization capital to revenue: (a) the contribution common to all firms (b_0t), and consequently is available to all firms (e.g., a certain level of employee education, the prevailing legal and institutional setting, etc.), and (b) a firm-specific contribution of organization capital to revenue (b_0stlog(SGA_it/SGA_i,t−1)) which is developed and enhanced by each firm (e.g., coded knowledge, production blueprints, business processes and procedures, marketing networks and channels, etc.). That is, the total factor productivity, a_0it, is firm-specific and influenced by SGA_it (as in equation (2)).

We estimate equation (1) by taking logarithms of annual changes, after substituting equation (2) into (1):

(3)

Equation (3) is estimated annually and cross-sectionally for each industry. The coefficient estimates indicate the average contributions of organization capital to revenue growth. In other words, these coefficient estimates provide us with the industry average efficiency with which companies use the resources to generate revenues.

We then use the coefficient estimates from equation (3) to obtain a firm-specific monetary measure of the contribution of organization capital to revenues. For each company in a given year, equation (3) with estimated coefficients allows us to predict the revenues under the average efficiency assumption without organization capital. We then subtract the predicted firm's revenues without firm-specific organization capital from the firm's actual revenues to get AbSALE_it, which is the contribution of organization capital to revenue of firm i in year t.

Similar to the above computation of the contribution of organization capital to revenues, we compute the contribution of organization capital to cost containment, and measure the latter by first predicting the costs based on the industry average cost of the resources utilized to achieve the sales. Then, the difference between the predicted cost and the actual cost, AbCOST_it, is the contribution of organization capital to cost containment.

The contribution of organization capital to operating profits is thus given by AbProfit_it= AbSALE_it+ AbCOST_it. The contribution of organization capital to revenues (AbSALE_it) and to cost containment (AbCOST_it) yields the firm a comparative advantage over a period of time. In other words, organization capital is not fickle in the sense that it contributes to AbSALE_it and AbCOST_it in one year but not for the next. Viewed differently, AbSALE_it and AbCOST_it are not due to a company being lucky or unlucky in a particular year. To account for these aspects, we capitalize and amortize AbProfit_it over five years to get our estimate of organization capital from the annual contributions. We scale the capitalized contributions of organization capital by the total assets in year t, which is referred as OC_it (organization capital).

Equations (3) and its cost counterpart are estimated each year for each of the twelve industry categories specified by Fama–French. The sample consists of all firms with sales and total assets greater than $5 million from 1971 to 2006. Data are obtained from the Compustat Annual Database. For stock returns tests, we obtain data from CRSP. The sample consists of 147,524 firm-year observations. The composition of sample by the twelve industry groups is provided in Panel A of Table 1.

Table 1.
MEASURE OF ORGANIZATION CAPITAL

Panel A: Industry distribution
Industry name	Total number of observations
Consumer nondurables	11,697
Consumer durables	5,079
Manufacturing	22,487
Oil, gas, and coal extraction and products	5,852
Chemicals and allied products	3,910
Business equipment	23,197
Telephone and television transmission	3,652
Utilities	7,469
Wholesale, retail, and some services	18,076
Healthcare, medical equipment, and drugs	9,179
Finance	17,701
Other	19,225
Total	147,524

Panel B: Descriptive statistics of variables used to estimate organization capital
	M	SD	Median	Q1	Q3
SALEit ($ millions)	1542.27	7289.60	150.25	43.10	632.71
COSTit ($ millions)	1281.96	6209.79	127.58	37.93	525.94
EMPit (hundreds)	8.82	32.55	1.38	0.37	5.05
PPEit ($ millions)	659.59	3299.62	32.62	7.42	188.84
SGAit ($ millions)	414.57	2140.60	35.25	8.64	143.74
Log (SALEit/SALEit−1)	0.1101	0.2459	0.0994	0.0017	0.2103
Log (COSTit/COSTit−1)	0.1132	0.2396	0.1006	0.0052	0.2110
Log (EMPit/EMPit−1)	0.0505	0.2554	0.0278	−0.0422	0.1289
Log (PPEit/PPEit−1)	0.0996	0.2884	0.0632	−0.0309	0.1995
Log (SGAit/SGAit−1)	0.1558	0.2581	0.1131	0.0328	0.2302

Panel C: Descriptive statistics of organization capital and firm performance
Variablea	M	SD	Median	Q1	Q3
Organizational capital variables
OCit	0.1439	0.2471	0.1118	0.0201	0.2382
OCit ($ millions)	246.63	1370.35	42.55	5.69	179.95
Operating performance and market performance
OIGrowthit+1	0.0217	0.0548	0.0149	−0.0035	0.0421
SALEGrowthit+1	0.1368	0.2219	0.0992	0.0175	0.2066
ABRETit	0.0306	0.3929	−0.0288	−0.2002	0.1764
RETit	0.1951	0.4726	0.1267	−0.0889	0.3774
Firm characteristics
SIZEit	6.0014	1.8647	5.9412	4.6561	7.2302
DIVit	0.0148	0.0179	0.0092	0.0000	0.0244
RDCAPit	0.1087	0.1537	0.0689	0.0307	0.1343
EPit	0.0769	0.0579	0.0674	0.0410	0.1014
D_EPit	0.0755	0.2642	0.0000	0.0000	0.0000
BMit	0.6973	0.4917	0.5922	0.3681	0.8888
BETAit	1.0543	0.5130	1.0611	0.7269	1.3108
VOLit	0.0151	0.0216	0.0117	0.0065	0.0182

• ^a   Variable definitions are provided in the Appendix.

Panel B of Table 1 provides the descriptive statistics of the variables used to estimate equation (3). The mean (median) SALE_it and COST_it are $1,542 ($150) and $1,281 ($127) million, respectively, indicating that the sample consists of many small firms and the empirical distribution is skewed to the right. Correspondingly, the resources also exhibit this skewness: the mean (median) PPE_it and SGA_it are $659 ($32) and $414 ($35) million, respectively; and the mean (median) EMP_it is 8.82 (1.38) thousand. The mean growth in SALE_it is roughly 11.64 per cent [=exp(0.1101) − 1] and COST is 11.98 per cent. It is interesting to note that on average the SALE_it and COST_it growth are similar. However, in the left tail of the distribution the sale growth is much lower than the cost growth, and in the right tail the sale growth is higher than the cost growth. On average, the growth in resources, that is, capital and labour, is lower than of sale growth and cost growth, while the growth in the capitalized selling, general and administrative expenses is much higher than the sale growth and cost growth, because selling, general and administrative expenses contains resources that influence organization capital. This likely indicates the importance of the role of intangibles.

The Dell Example

Since the contribution of organization capital is derived from the firm's performance, the question arises whether the organization capital measure adds useful information beyond that in typical financial data such as sales and net income. This question is addressed comprehensively in the next section, but the following case of Dell Corp., which fell on hard times in 2005, provides a rough idea. Figure 1 presents the sales, net income, organization capital and stock price data for Dell from 2001 to 2007. We scale all these variables by their respective levels as of 2001. Thus, stock price in July of 2005 is the stock price of July divided by the stock price as of January of 2001. It follows that all variables begin with a value of 1, so as to highlight the relative trends of these variables over the period. As shown in Figure 1, sales, net income and stock prices exhibit similar trends from 2002 to 2005. In fact they track each other quite consistently, even though the stock price is a forward looking measure while net income and sales are backward looking numbers. Nevertheless the trends exhibited by the accounting-based and stock-based performance measures are similar. Starting in 2005, the stock price starts to decline precipitously up until the beginning of 2006, due to serious concerns over the continued growth of Dell, its governance practices and accounting irregularities.

Interestingly, the organization capital measure shows a different trend. Starting from 2001, the organization capital measure drops, up until 2004, and flattens out for 2005. This is starkly different from the backward looking sales and net income variables, which exhibit an increasing trend throughout the period. This drop in organization capital corresponds to the coming to light of the slowing growth and the consequent accounting irregularities in 2005/2006.³ That is, the ‘real’ performance of Dell was not as good as the accounting numbers reflected; when this came to light as a ‘surprise’ in 2005/2006 the stock price dropped. The organization capital measure, at least in Dell's case, provided an early warning of the real performance. We now generalize and examine the relationship between organization capital and future firm performance for the entire sample.

RELATING ORGANIZATION CAPITAL TO FUTURE PERFORMANCE

We relate the organization capital measure (OC) to the five future years performance. We use growth of operating income (OIGrowth_i,t+i) and growth of sales (SALEGrowth_i,t+i) as measures of operating performance, and use the size and book-to-market adjusted annual return and raw annual return as measures of market performance. Operating income growth (OIGrowth_i,t+i) is defined as average operating income for years t+1 and t+i, minus the operating income in year t scaled by total assets in year t. Sales growth (SALEGrowth_i,t+i) is computed as average sales for years t+1 and t+i, minus the sales in year t scaled by sales in year t. We compute the size and book-to-market adjusted excess returns using the companion portfolio approach, where firms are grouped by the book-to-market ratio into five equal groups at the end of June each year, and the size breakpoints are determined by classifying the NYSE companies into five equal groups in June each year. Thus, we have five groups of book-to-market ratio and five groups of firm size to determine the companion portfolio. The monthly excess returns are then computed as the difference between the firm's monthly return and the companion portfolio's monthly return. The annual excess returns are obtained by compounding the monthly excess returns each fiscal year. For this analysis we require that firms have an OC (organization capital) estimate as well as the performance data be available for the next five years. The final sample consists of 27,701 firm-year observations.

Table 1, Panel C provides the descriptive statistics of organization capital and firm performance. The mean (median) organization capital is 0.1439 (0.1118), indicating that on average organization capital represents roughly 14 per cent of the total assets. In unreported analysis, the mean (median) contributions of organization capital to sales and cost containment are 0.2185 and −0.0745 (0.1740 and −0.0568), respectively. The contribution of organization capital through cost containment is negative for roughly 80 per cent of the firm-year observations. This suggests that, in general, organization capital enables firms to achieve additional output, and presumably firms have to ‘pay extra’ in terms of cost to get the additional contribution to output.

Table 1, Panel C also provides descriptive statistics of other variables used in the multivariate analysis. The mean (median) firm-size measured as the logarithm of market value of equity is 6.0014 (5.9412), which corresponds to a market value of equity of $404 ($380) million. The firms considered in our sample are reasonably large, because we require data for five prior years to compute OC and for five future years: a ten-year survival requirement. The mean (median) book-to-market ratio is 0.6973 (0.5922), suggesting that the sample firms are on average high growth firms. Roughly 8 per cent of the sample contains firms with negative bottom-line earnings: mean D_EP_it is 0.0755. Overall, the firm characteristics indicate that the sample contains reasonably mature firms with reasonable growth potential.

Organization Capital and Operating Performance

If the organization capital measure captures an attribute of efficiency with which firms convert their resources into outputs, then this capability should result in superior firm performance in future years. In other words, if organization capital is simply a backward-looking measure that does not capture the attribute of differential efficiency that can be sustained in the future, then it should not be systematically associated with the future operating performance. To examine this, we form ten portfolios based on the OC measure. In particular, we classify each firm into the respective industry-year, decile based on OC: OC is the capitalized abnormal operating profit scaled by total assets. We then track the operating performance measures for the five years subsequent to the year of portfolio formation.

Table 2, Panel A provides the mean of cumulative growth in operating income (OIGrowth_it+i) for each OC group for one, two, three, four and five years following the year of portfolio formation. Firms in OC group 1 (10) are the firms in the lowest (highest) decile in their industry for that year. The mean OIGrowth_it+i of firms in the lowest (highest) OC group for one, two, three, four and five years ahead are: 2.20 per cent, 3.32 per cent, 4.35 per cent, 5.38 per cent and 6.47 per cent (3.53 per cent, 5.32 per cent, 7.34 per cent, 9.65 per cent and 12.07 per cent), respectively. The last two rows show the difference between the top and bottom OC groups and the related t-statistic. While the difference is positive and statistically significant for all five years, there is a steady increase in the difference. This increase is attributable to a more marked increase in the top OC group of about 2.41 [=(12.07 − 3.53)/3.53] times compared to an increase in the bottom group of about 1.94 [=(6.47 − 2.20)/2.20] times. In unreported analysis, for the one-year ahead OI growth, the percentage of negative OI growth is roughly 30 per cent in the bottom OC group and 20 per cent in the top OC group; while for the cumulative five-year ahead OI growth the percentage of negative OI growth is roughly 28 per cent and 23 per cent. This indicates that, on average, firms in the bottom OC group appear to exhibit a low propensity to grow their operating income.

Table 2.
ORGANIZATION CAPITAL AND FUTURE OPERATING PERFORMANCE, UNIVARIATE TESTS

Panel A: Operating performance is OIGrowth it +i
Portfolio of OCit	i is years after portfolio formation
	i= 1	i= 2	i= 3	i= 4	i= 5
1: Bottom	0.0220	0.0332	0.0435	0.0538	0.0647
2	0.0166	0.0259	0.0349	0.0438	0.0524
3	0.0154	0.0245	0.0333	0.0422	0.0513
4	0.0160	0.0247	0.0335	0.0428	0.0528
5	0.0190	0.0294	0.0401	0.0511	0.0631
6	0.0194	0.0302	0.0407	0.0520	0.0641
7	0.0217	0.0336	0.0472	0.0614	0.0757
8	0.0248	0.0381	0.0520	0.0671	0.0832
9	0.0277	0.0437	0.0614	0.0810	0.1016
10: Top	0.0353	0.0532	0.0734	0.0965	0.1207
Top minus Bottom	0.0134*	0.0200*	0.0299*	0.0427*	0.0560*
t-value	7.32	8.45	10.2	11.91	13.02

Panel B: Operating performance is SALEGrowth it +i
Portfolio of OCit	i is years after portfolio formation
	i= 1	i= 2	i= 3	i= 4	i= 5
1: Bottom	0.1228	0.1901	0.2556	0.3264	0.4033
2	0.1109	0.1731	0.2365	0.3007	0.3677
3	0.1181	0.1846	0.2504	0.3190	0.3937
4	0.1194	0.1841	0.2533	0.3255	0.4036
5	0.1272	0.1974	0.2734	0.3525	0.4362
6	0.1323	0.2065	0.2819	0.3629	0.4467
7	0.1429	0.2241	0.3138	0.4075	0.5045
8	0.1529	0.2330	0.3196	0.4151	0.5163
9	0.1627	0.2598	0.3665	0.4808	0.6013
10: Top	0.1793	0.2849	0.3986	0.5257	0.6615
Top minus Bottom	0.0565*	0.0948*	0.1430*	0.1993*	0.2582*
t-value	8.12	9.55	10.94	11.9	12.4

• Note: Variable definitions are provided in the Appendix.

Table 2, Panel B provides the mean sales growth (SALEGrowth_it+i): sales growth shows a similar trend to income: firms in the top OC group exhibit a higher sales growth than firms in the bottom OC group throughout the future five years. Figure 2 portrays the difference in IOGrowth_it+1 and SALEGrowth_it+1 across the top 30 per cent and bottom 30 per cent of the OC groups, and shows that the difference is substantial. Overall, these results show that organization capital is associated with future operating performance of firms.

The univariate tests presented in Table 2 have the advantage of not having a functional form. However, to the extent that OC measure is associated with other firm characteristics, such as firm size (e.g., smaller firms having more OC or vice versa), the relationships in the univariate tests may not be conclusive. We augment Lev and Nissim's (2004) model by adding the rank of OC and D_EP (Penman and Zhang, 2002) to the analysis. In particular, we estimate the following equation:

(4)

where Growth = {OIGrowth_it+1, SALEGrowth_it+1}. OIGrowth_it+i is the average difference between operating income in year t+i minus operating income in year t, scaled by total assets in year t. SALEGrowth_it+i is the average difference between sales in year t+i minus operating income in year t, scaled by sales in year t. R_OC_it is the industry-year based decile rank of OC_it, and OC_it is the contribution of organization capital to revenues and cost containment. R_OC_it is scaled to be between 0 and 1. SIZE_it is the natural log of market value of equity. DIV_it is dividend to common shares, scaled by total assets. RDCAP_it is the sum of R&D expenditures and capital expenditures, scaled by sales. EP_it is net income divided by market value of equity if the ratio is greater than 0, and 0 otherwise. D_EP_it is an indicator variable that equals one if net income divided by market value of equity is less than 0, and zero otherwise. BM_it is the book value of equity divided by market value of equity.

Table 3, Panel A provides the mean coefficient estimates obtained from the annual estimation of equation (4), where the t-statistics are based on the standard errors of the annual coefficient estimates, that is, the Fama–MacBeth procedure (see Fama and MacBeth, 1972). The coefficient estimates on R_OC_it for one, two, three, four and five years ahead of operating income growth are 0.0074, 0.0113, 0.0177, 0.0263 and 0.0355, respectively, and all are statistically significant. These coefficient estimates indicate the difference between the top and bottom OC groups, because R_OC_it is scaled to be between 0 and 1. The results indicate that, after controlling for other major factors that are associated with future operating performance, organization capital still contributes to future growth in operating income. Compared to the univariate analysis, the difference across the top and bottom OC groups is smaller in magnitude in the multivariate tests (Table 3) than the univariate tests (see Table 2), suggesting that the other factors that influence growth are also correlated with the organization capital, yet don't subsume it.

Table 3.
ORGANIZATION CAPITAL AND FUTURE OPERATING PERFORMANCE, MULTIVARIATE TESTS

Panel A: Dependent variable is OIGrowth it +i
	i is years after portfolio formation
	i= 1		i= 2		i= 3		i= 4		i= 5
	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value
INTERCEPT	0.0598*	13.12	0.0899*	13.98	0.1193*	14.23	0.1487*	13.65	0.1780*	13.45
R_OCit	0.0074*	4.17	0.0113*	4.83	0.0177*	5.83	0.0263*	6.96	0.0355*	7.77
SIZEit	−0.0027*	−6.85	−0.0044*	−8.39	−0.0061*	−9.61	−0.0080*	−9.82	−0.0100*	−9.73
DIVit	−0.2181*	−6.90	−0.3226*	−7.01	−0.4434*	−6.94	−0.5516*	−6.62	−0.6509*	−6.37
RDCAPit	0.0176*	3.16	0.0315*	4.82	0.0448*	6.98	0.0606*	8.99	0.0786*	9.58
EPit	−0.1435*	−8.53	−0.2044*	−9.23	−0.2567*	−9.42	−0.3020*	−8.94	−0.3428*	−8.33
D_EPit	0.0112*	5.02	0.0085*	2.95	0.0064	1.88	0.0043	1.10	0.0028	0.63
BMit	−0.0233*	−9.00	−0.0322*	−8.72	−0.0412*	−8.69	−0.0510*	−8.50	−0.0614*	−8.48
Adj. R2	0.0682		0.0853		0.0988		0.1074		0.1118

Panel B: Dependent variable is SALEGrowth t +i
	i is years after portfolio formation
	i= 1		i= 2		i= 3		i= 4		i= 5
	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value
INTERCEPT	0.2998*	17.26	0.4646*	18.03	0.6355*	18.95	0.8143*	19.53	1.0029*	20.27
R_OCit	0.0090	1.08	0.0222	1.82	0.0457*	2.83	0.0778*	3.71	0.1123*	4.39
SIZEit	−0.0091*	−6.21	−0.0160*	−7.42	−0.0239*	−8.45	−0.0330*	−9.33	−0.0436*	−10.15
DIVit	−2.1746*	−14.41	−3.2522*	−14.07	−4.4270*	−13.32	−5.6400*	−12.86	−6.8185*	−12.57
RDCAPit	0.1401*	4.67	0.2389*	6.47	0.3343*	8.07	0.4428*	9.04	0.5657*	9.67
EPit	−0.3934*	−5.51	−0.5825*	−5.44	−0.7611*	−5.14	−0.9269*	−4.95	−1.0929*	−4.84
D_EPit	−0.0666*	−9.10	−0.0976*	−8.13	−0.1221*	−6.60	−0.1438*	−6.14	−0.1590*	−5.45
BMit	−0.1032*	−10.82	−0.1524*	−11.52	−0.2032*	−11.98	−0.2549*	−12.16	−0.3079*	−12.23
Adj. R2	0.0838		0.0995		0.1102		0.1169		0.1202

• *  Significant at the 5% level.
• Note: The coefficient estimates are the mean of the annual regressions; the t-statistics are based on the standard errors of the annual coefficient estimates, and variable definitions are provided in the Appendix.

Table 3, Panel B provides the results of estimating equation (4) when the dependant variable is sales growth (SALEGrowth_it+i). The first two years OC is not associated with sales growth after controlling for the other factors. However, over three years ahead OC is significant and exhibits and increasing trend, similar to the OI growth. Overall, the results are qualitatively similar to those of Panel A. In summary, our OC estimate is associated with future operating performance, an indication that the organization capital measure captures fundamental efficiency attributes affecting long-term performance.

Organization Capital and Future Stock Returns

The results in Tables 2 and 3 show that our measure of organization capital is associated with future operating performance. So, OC, as the Dell example shows, is forward looking. The natural question is: Do investors fully understand and incorporate the information in organization capital? In other words, does the stock return incorporate contemporaneously the elements of OC? These are important questions, because hardly any useful information is provided by firms to investors about intangibles in general, or about organization capital, which is an aggregator of most intangibles, in particular.

For this purpose, similar to the analysis in Table 2, we track the mean cumulative excess stock returns of the ten OC groups in the five subsequent years. For excess returns, ABRET, we use the size and book-to-market adjusted annual returns. The two years ahead cumulative excess returns, CUMABRET_it+i, is the sum of ABRET for years t+1 and t+2; similarly, the five year ahead cumulative excess returns is the sum of ABRET from years t+1 to t+5.

Table 4, Panel A shows the cumulative excess returns, CUMABRET_it+i, for the ten OC deciles from years t+1 to t+5. The difference in cumulative excess returns between the top and bottom OC groups for one year ahead are: 3.53 per cent, 6.49 per cent, 8.33 per cent, 10.12 per cent and 11.54 per cent, respectively. For the fourth and fifth years the difference in cumulative excess returns across the top and bottom OC groups flattens out considerably. This occurs because the bottom OC groups' cumulative return flattens out while the top OC group exhibits a considerably lower excess returns than in the first two years. We will examine this reversal of excess returns, that is, the flattening out of the cumulative excess returns in Table 4, Panel B. It is also interesting to note that the cumulative excess returns for groups 2, 3 and 4 are markedly lower than those of groups 7, 8 and 9, and all of these intermediate groups' cumulative excess returns also flatten out in the fourth and fifth years. The excess returns for the top and bottom groups are large compared to the intermediate groups, even in the fifth year: this likely indicates that some of the excess return is likely attributable to inadequate adjustment for risk. While for the bottom OC group the risk is likely to be primarily associated with business risk, for the top OC group the risk is likely to be mainly attributable to poor information, that is, information risk. Figure 3 shows the difference in CUMABRET_it+i across the top 30 per cent and the bottom 30 per cent of the OC groups for the five subsequent years: the results are similar to those of the top and bottom 10 per cent of the OC groups.

Table 4.
ORGANIZATIONAL CAPITAL AND FUTURE EXCESS RETURNS

Panel A: Future cumulative excess returns, CUMABRETit+i
Portfolio of OCit	i is years after portfolio formation
	i= 1	i= 2	i= 3	i= 4	i= 5
1: Bottom	0.0332	0.0576	0.0860	0.1038	0.1170
2	0.0222	0.0504	0.0583	0.0615	0.0607
3	0.0123	0.0278	0.0247	0.0258	0.0232
4	0.0175	0.0313	0.0443	0.0483	0.0461
5	0.0277	0.0526	0.0582	0.0818	0.0834
6	0.0304	0.0604	0.0711	0.0817	0.0870
7	0.0211	0.0478	0.0769	0.0917	0.0921
8	0.0230	0.0371	0.0636	0.0768	0.0918
9	0.0517	0.0957	0.1248	0.1580	0.1713
10: Top	0.0685	0.1225	0.1693	0.2050	0.2325
Top minus Bottom	0.0353*	0.0649*	0.0833*	0.1012*	0.1154*
t-value	2.85	3.77	4.02	4.31	4.44

Panel B: Future annual excess returns, ABRETit
Portfolio of OCit	i is years after portfolio formation
	i= 1	i= 2	i= 3	i= 4	i= 5
1: Bottom	0.0332	0.0244	0.0284	0.0178	0.0132
2	0.0222	0.0282	0.0079	0.0032	−0.0008
3	0.0123	0.0155	−0.0031	0.0011	−0.0025
4	0.0175	0.0138	0.0130	0.0040	−0.0022
5	0.0277	0.0249	0.0057	0.0236	0.0015
6	0.0304	0.0300	0.0107	0.0106	0.0053
7	0.0211	0.0267	0.0292	0.0147	0.0004
8	0.0230	0.0141	0.0265	0.0132	0.0150
9	0.0517	0.0440	0.0291	0.0331	0.0134
10: Top	0.0685	0.0540	0.0467	0.0357	0.0275
Top minus Bottom	0.0353*	0.0295*	0.0184	0.0179	0.0143
t-value	2.85	2.49	1.59	1.58	1.27

Panel C: Dependant variable is CUMABRETt+i a
	i is years after portfolio formation
	i= 1		i= 2		i= 3		i= 4		i= 5
	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value	Coef.	t-value
INTERCEPT	0.0751	1.84	0.1112*	2.22	0.1413*	2.36	0.1768*	2.84	0.1458*	1.99
R_OCit	0.0299*	2.93	0.0509*	3.44	0.0750*	4.08	0.0925*	4.54	0.1100*	4.67
SIZEit	−0.0169*	−5.39	−0.0294*	−6.98	−0.0385*	−7.55	−0.0460*	−7.90	−0.0477*	−7.77
BETAit	0.0118	0.62	0.0248	1.22	0.0349	1.45	0.0328	1.09	0.0466	1.35
VOLit	1.0212*	3.01	2.4292*	4.74	3.3668*	6.10	4.3802*	6.03	5.0360*	5.82
EPit	0.1522	1.00	0.1727	1.03	0.1218	0.69	0.1773	0.89	0.1195	0.58
D_EPit	0.0815*	3.33	0.1205*	3.50	0.1421*	4.25	0.1623*	4.57	0.1773*	4.37
BMit	0.0010	0.08	0.0183	1.11	0.0311	1.54	0.0246	1.18	0.0413	1.59
Adj. R2	0.0124		0.0201		0.0269		0.0293		0.0309

• ^a   In panel C, the coefficient estimates are the mean of the annual regressions and the t-statistics are based on the standard errors of the annual coefficient estimates.
• Note: Variable definitions are provided in the Appendix.

Table 4, Panel B provides the ABRET_it for the ten OC groups to examine the reversal of hedge returns. The difference in future excess returns across the top and bottom OC groups, one and two years after portfolio formation, are: 3.53 per cent and 2.95 per cent, and is zero for three, four and five years ahead. This is similar to the cumulative excess returns shown in Panel A. We show the annual excess returns to highlight the reversion of hedge returns, that is, the difference across the top and bottom OC groups tends to zero. This is important, since the reversal of the excess returns indicates that the difference in excess returns across OC groups is attributable to mispricing, which is likely to occur due to lack of information on the intangibles, rather than to inappropriate risk adjustment in our analysis.

We also test the relationship between OC and future cumulative excess returns with multivariate regressions. We augment the Lev and Nissim (2004) model by adding the rank of OC and D_EP. Accordingly, we run Fama–MacBeth regression for the following equation:

(5)

where CUMABRET_it+1 is sum of excess return adjusted for the companion size and book-to-market from year t to year t+i. R_OC_it is the scaled decile rank of OC_it. SIZE_it is the natural log of market value of equity. BETA_it is the slope coefficient obtained from estimating a market model using the previous sixty monthly returns. VOL_it is the firm-specific variance of the monthly returns computed over the previous sixty months. EP_it is net income divided by market value of equity if the ratio is greater than zero, and 0 otherwise. D_EP_it is an indicator variable that equals one if net income divided by market value of equity is less than 0, and zero otherwise. BM_t is the book value of equity divided by market value of equity.

Table 4, Panel C provides the results of estimating equation (5). The mean coefficient estimates obtained from the annual estimation of equation (5) and the t-statistics are computed using the standard errors of the annual coefficient estimates, that is, the Fama–MacBeth procedure (see Fama and MacBeth, 1972). The coefficient estimate on R_OC_it indicates the difference between the top and bottom OC groups, controlling for other factors. In particular, the coefficient estimates on R_OC_it for the next five years are 2.99 per cent, 5.09 per cent, 7.50 per cent, 9.25 per cent, and 11.00 per cent, respectively, indicating a trend of flattening out but not as markedly as in the univariate analysis discussed above.

Summarizing, the tests document a systematic association between our estimate of firm-specific organization capital, OC, and future accounting-based performance measures as well as market-based measures. While these results support the validity of the OC measure as reflecting fundamental efficiency dimensions, the implications of the two tests are different. In particular, the stock return tests indicate that despite the importance of the OC indicator, investors don't fully comprehend this information in a timely manner. The OC measure, therefore, both reflects important operating dimensions and is largely unknown to investors.

We next proceed to examine the association between executive compensation and organization capital to corroborate the attribute of organization capital as a measure of management quality.

RELATING ORGANIZATION CAPITAL TO EXECUTIVE COMPENSATION

Our findings thus far indicate that the measure of organization capital, which is an aggregate measure of intangibles, is related to firms' future operating and stock performance. In this section, we examine whether the organization capital can be used as a measure of managerial quality for decisions such as executive compensation.

Our measure of organization capital reflects the aggregate value of management quality, that is, the intangibles that enable physical and human resources to be used more efficiently. As such, the measure of organization capital should be associated with the ability of management, and if executive compensation is a reward to managerial ability, then the measure of organization capital should be associated with executive compensation. Accordingly, we examine the association between executive compensation and organization capital.

For this purpose, we use two different measures of CEO compensation: total compensation scaled by firms' total assets, and the pay-for-performance sensitivity (PPS) of the executives' equity holdings. Total compensation is the sum of salary, bonus, other annual benefits, restricted stock grants, long-term incentive plan (LTIP) payouts, options grants based on Black–Scholes value and all other benefits. We obtain data for total compensation from EXECOMP database.

We estimate the pay-for-performance sensitivity of equity (PPS) using the procedure of Core and Guay (1999). PPS is defined as the change in the dollar value of the CEO's stock and options for a 1 per cent change in the stock price. While computing this measure is straightforward for stockholdings, because stock value increases by 1 per cent for each 1 per cent increase in the stock price, computing this measure for options is not as straightforward, because the percentage increase in the value of an option is less than the percentage increase in the stock price, and depends upon the parameters embedded in the option contract. Following Core and Guay, we estimate PPS of options as the partial derivative of one option's value with respect to stock price (the option delta), and then multiply the option delta by 1 per cent of the firm's stock price. Thus, we estimate the incentives from CEO's option portfolio as the sum of the option deltas multiplied by 1 per cent of price; and PPS is the sum of the incentives from the stock holdings and option holdings. We obtain data on CEOs' equity holding and accumulated option grants from EXECOMP database. In the empirical analysis we use the annual industry median as the benchmark to adjust the total compensation and PPS measures, so as to account for industry-specific variation.

Total compensation is the annual ‘flow’ of compensation, and thus captures the reward to CEOs based on their ability. The PPS measures the exposure of CEOs to the stock price movements based on the CEOs' current and potential equity holding. PPS should also be influenced by the CEOs' ability because higher ability CEOs are more likely to capture more rents. In other words, higher ability CEOs are likely to have more risk in their current and potential equity portfolios, and thus as a compensation for risk the corresponding reward is also likely to be higher. The PPS measure is a stock of reward, that is, all outstanding equity grants from pervious periods are accumulated in computing PPS. As such, PPS measures the exposure of the CEO to the firm's stock price.

We obtain the CEOs' total compensation and variables required to estimate PPS from EXECOMP database from 1992 to 2006. The data required to calculate firm size, sales growth and firm's market share are obtained from Compustat. The variable of corporate governance index (G-Index) is obtained from Corporate Governance of IRRC database, and is based on the measure developed by Gompers et al. (2003). This index is comprised of twenty-four indicators reflecting the shareholder rights and is reverse ordered, that is, higher index values indicate weaker shareholder rights. While the G-Index measure is available for 1990, 1993, 1995, 1998, 2000, 2002 and 2004, we use the latest available G-Index value for the years where data are not available. We adjust firm size, sales growth and G-Index with the annual industry median as the benchmark so as to account for industry practices and factors.

We examine the relationship between our organization capital measure in year t+1 and executive compensation in year t. We do this because compensation presumably motivates the CEO to direct his/her attention to generate organization capital in future years. If organization capital is the abnormal contribution to operating profits arising from the conglomeration of intangibles of which managerial ability is key, then the compensation in year t must motivate managers to get the abnormal contribution to profits in future years.⁴ Our test design captures this notion of rewards/incentives embedded in executive compensation.

Panel A of Table 5 provides descriptive statistics of the variables used in the analyses. The mean of OC_it+1 is 0.0733, indicating that firms in general have positive organization capital. The mean (median) of CEO's total compensation scaled by firm's total asset and CEO's PPS are 2.4873 (1.2329) and 782.96 (188.48), indicating that the executive compensation variables are right skewed.

To examine the association between executive compensation and organization capital, we form five portfolios each year based on the industry adjusted measures of executive compensation. We then track the organization capital of each portfolio in years t and t+1. Panel B of Table 5 provides the results of the portfolio based on the total compensation scaled by total assets in year t. The difference in the organization capital between the top and bottom groups are 0.049 (t-value = 7.14) and 0.035 (t-value = 4.87), in year t and t+1, respectively. This indicates that higher executive compensation is associated with higher subsequent organization capital. Panel C of Table 5 provides the results of the portfolio based on pay-performance-sensitivity of equity holdings: The results are qualitatively similar to those obtained with levels of executive compensation. Overall, these results are consistent with the notion that managers with higher compensation have higher ability as reflected by higher organization capital.

Organization capital can be attributed to other intangibles that may not be related to manger's ability/management quality. In other words, organization capital may be associated with firm characteristics, such as firm size, sales growth, market share and corporate governance, the same characteristics that are associated with executive compensation. Thus, we need to control for these factors to focus on the association between executive compensation and organization capital. For this purpose, we estimate the following equation.

(6)

where OC_it+1 is the organization capital in year t+1. EXECOMP_it= {R_TC_it, R_PPS_it}. R_TC_it is the industry-year based quintile rank of total compensation scaled by total assets in year t. R_PPS_it is the industry-year based quintile rank of pay-performance-sensitivity of holding equity (Core and Guay, 1999) in year t. ADJ_SIZE_it is industry-year median-adjusted firm size. ADJ_ SALEGROW_it is industry-year median-adjusted sales growth. SHARE_it is market share of a firm, equal to firm's sale divided by total industry sales in year t. ADJ_GINDEX_it is industry-year median adjusted corporate governance index (G-index).

Table 6, Panel A reports the results of estimating equation (6) with R_TC_it. The coeffcient on R_TC_it is 0.0623 (t-value = 6.46) in column (1) and is 0.0616 (t-value = 6.40) in column (2), indicating that after controling for other factors that are associated with compensation, OC is still associated with compensation. The results in Panel B using R_PPS_it also provide qualitatively similar results.

CONCLUDING REMARKS

Intangibles or knowledge assets are major drivers of corporate and national growth. Organization capital enables productive interactions among various resources (both tangible and intangible) for creating economic value and growth. Organization capital—a major form of intangibles, embodied in unique organizational designs and processes—is the least documented type of intangible assets. We develop a measure for a company's organization capital, incorporating a firm's potential to generate super-normal revenues and cost savings. We validate the measure by examining its association with future performance measures, such as operating income growth, sales growth and abnormal returns. We find that the organization capital measure is associated with five years of future operating and stock return performance, after controlling for other factors. Thus, our organization capital measure captures fundamental efficiency dimensions of operations.

We also examine the association of the organization capital measure with executive compensation. Executive compensation is an incentive mechanism as well as a measure of the executive's ability that should manifest in business processes and systems, that is, the way of doing business. We find that executive compensation is positively associated with our measure of organization capital.

Collectively the results show that organization capital is an important intangible asset that is related to firm value and important corporate decisions. While we have developed the measure from a firm-level aggregate set of input and output measures, providing supplementary disclosure on inputs and outputs at a segment or divisional level would help to improve the measurement and tracking of the firm's important intangible asset, its organization capital. In general, evidence exists that investment in intangibles, and organization capital in particular, is relevant for investors and managers. However, such information is disclosed in a haphazard manner, similar to the disclosure of segment reporting before it was standardized. Thus, it would be useful if standard setting bodies develop templates for disclosure of intangibles, such as inputs and outputs, so that it can be disseminated in a standardized fashion.