Does Implementing an Effective TQM Program Actually Improve Operating Performance? Empirical Evidence From Firms that have Won Quality Awards

This study explores the hypotheses that implementing effective TQM programs improves the operating performance of firms. The winning of quality awards is used as a proxy for the effective implementation of TQM programs. Changes in various performance measures for a test sample of quality award winners are compared against a sample of control firms. Our statistical tests provide strong evidence that firms that have won quality awards outperform the control firms on operating income based measures. Over a ten-year period, starting six years before to three years after the year of winning the first quality award, the mean (median) change in the operating income for the test sample is 79% (30%) higher than that of the control sample. There is reasonably strong evidence that firms that have won quality awards do better on sales growth than the control firms. Over the ten-year period, the mean (median) change in sales for the test sample is 43% (18%) higher than that of the control sample. We also find weak evidence that firms in our test sample are more successful in controlling costs when compared to the firms in the control sample.

In addition, the results indicate that firms in our test sample increased their capital expenditures more than the control sample over the time period prior to winning quality awards. Compared to the control sample, the test sample shows higher growth in both employment and total assets. The inventory turnover performance of the test firms is not different from that of the control sample.

1. Introduction

Recently, concerns have been raised about whether Total Quality Management (TQM) programs have generated real economic gains and/or improvements in operating performance. Recent studies by some management consultancy firms suggest that many TQM programs have not been effective (see for example the International Quality Study, a joint project between Ernst & Young and American Quality Foundation [1992], Rath and Strong [1992], and Kelly [1992] which discusses the study done by Arthur D. Little Inc. ). These studies have received publicity in the business press with headlines such as “The Cost of Quality: Faced with Hard Times, Business Sours on Total Quality Management” (Mathews and Katel [1992]), “Total Quality is Termed Only Partial Success” (Fuchsberg [1992a]), “Quality Programs Show Shoddy Results” (Fuchsberg [1992b]), and “Why Most Quality Efforts Fail” (Szwergold [1992]). Furthermore, stories have appeared about some of the Malcolm Baldrige National Quality Award winners that have suffered financial setbacks, layoffs, and even bankruptcies. The conclusions often drawn from these stories and studies are that TQM is not as effective as previously believed, might even damage firm performance, and it is a fad that has run its course and is losing popularity.

Proponents of TQM have reacted to the negative publicity by reminding critics of the few well publicized success stories of TQM, and used these to make the case that TQM works. Others have revisited the theory of TQM to reiterate what is TQM, why there is nothing wrong with the theory of TQM, and what organizations need to do to successfully implement TQM (King [1992], Johnson [1993], Senge [1993], and Grant, Shani, and Krishnan [1994], among others). Others have indicated that TQM adoption rates are increasing not decreasing, suggesting that TQM is alive and well (Haim [1993]).

It is surprising and somewhat disturbing that claims and counter-claims about whether TQM programs have paid off in a financial sense are rarely supported by objective and rigorous empirical evidence. For example, the highly publicized International Quality Study did not provide any statistical data on the effectiveness of TQM programs, but reported that firms may waste millions of dollar on TQM strategies that do not improve performance and may even hurt performance. Similarly, there are studies which suggest that TQM improves operating performance but, with a few exceptions, rarely support it with statistical evidence (the next section briefly reviews these studies). The success and failure of TQM programs should be judged against results from rigorous empirical evidence. It seems to us that expectations about how TQM can improve performance are based less on rigorous empirical evidence, and more on anecdotes, hype, and publicity. Furthermore, the expectations seems to be very high.

This paper provides empirical evidence on whether effective implementation of TQM programs affects the operating performance of firms. The winning of quality awards is used as a proxy for the effective implementation of TQM programs. Our results are based on a sample of nearly 400 publicly traded firms who won their first quality award between 1983 and 1993. We use publicly available accounting data to test for changes in operating performance that result from implementing effective TQM programs. Operating performance is examined over a ten-year period starting six years before to three years after the year of winning the first quality award.

There are two reasons for using a sample of quality award winners. First, an important objective of quality award givers is to recognize firms that have done an outstanding job in implementing effective TQM programs. To maintain credibility and the value of awards, award givers have strong incentives to give awards to only those firms that have significantly improved quality in an effective manner. Award givers typically decide on winners after conducting an independent evaluation and assessment of a firm's quality practices and measuring a firm's quality performance against some preestablished standards. For a firm to win such an award, it must undergo a strict and effective quality improvement program. Therefore, winning a quality award by a firm provides independent third-party certification that the firm has implemented an effective TQM program. By focusing on quality award winners we avoid the biases associated with asking firms to self-judge the effectiveness of their TQM programs.

We note that the system of quality awards is a source of controversy. Given the popularity of the Baldrige Award, it is natural to find this award at the center of controversy (Garvin [1991]). The criticisms leveled against the Baldrige Award could very well apply to any other quality award. Both the critics and the supporters have passed judgment on the value of quality awards based on conceptual arguments and anecdotes, but not on rigorous empirical evidence.

The second reason for using a sample of quality awards in our study is that award givers exclude financial performance in the selection of award winners because of technical, fairness, and confidentiality considerations. Thus, firms in our sample are not selected because of their financial performance (good or bad). By basing the results of our study on publicly traded firms that have won quality awards, information on financial performance is obtained independently of TQM practices. Thus, we avoid the biases associated with asking firms to self-report the impact of TQM on financial performance.

Our study is one of the few attempts to estimate the long-term operating effects of implementing effective TQM programs. It fills a gap that exists in the literature on TQM. The results of our study provide a benchmark of what can be expected from an effective TQM program. It also helps, to some extent, in resolving the controversy about the magnitude of economic gains from TQM programs.

Section 2 summarizes the existing empirical evidence on the effect of TQM on financial performance. Section 3 briefly discusses the various hypotheses examined, and the performance variables used in this study. Section 4 describes the sample selection process and the methodology. Section 5 discusses the empirical results. The final section summarizes the paper and outlines future research directions.

2. Review of existing empirical evidence on TQM and financial performance.

The extent of existing evidence on TQM is perhaps best discussed by Haim [1993].9 He summarized the results of 20 different empirical studies on TQM. Most of these studies were conducted by business organizations and consulting firms using surveys. Of the 20 studies reviewed, 15 provide evidence on the impact of TQM and related practices on internal, external, and bottom-line measures. Twelve of these studies rely solely on the perceptions of managers, two add external validity by combining perceptions of respondents with analysis of company records, and one uses externally judged measures of both TQM and performance. Only three of the 20 studies report any kind of numerical measurement of the profitability impact of TQM (these studies are briefly reviewed next). Other studies simply give opinions about whether TQM improved the bottom line performance or not. Haim [1993] notes that in most studies the findings consist entirely of opinions of managers who completed the surveys, and are rarely based on objective data. Furthermore, there has been little in the way of independent measurement of TQM practices and their impact on financial or non-financial measures of performance.

The study by the United States General Accounting Office [1991] is one of the three studies which Haim [1993] indicated has any kind of numerical measurement of the impact of TQM on profitability. This study examined the effect of TQM practices on financial performance based on responses from 22 firms that were finalists or winners in the 1988 and 1989 Baldrige Award Competition. The study measured operating results using measures such as market share, sales per employee, return on sales, and return on assets. For the 15 firms that responded, the study found that 34 of the 40 observations increased and 6 declined. Responses were also favorable in the areas of customer satisfaction, quality, cost, and employee relations.

The second study is by Fitzerald and Erdmann [1992], who estimate the impact of continuous improvement practices, a key element of TQM. Based on responses received from more than 280 automotive suppliers, their survey shows that over a two to three year period, respondents reported an average of 17% increase in profits as a result of their continuous improvement efforts.

The third study is an internal study by International Business Machines (see Quality Management Update [1993]) which compared the performance of 57 business units that scored 500 or more out of 1000 on the Baldrige Criteria, with other business units that did not. The study found that these 57 units outperformed the other business units in areas of customer satisfaction, employee morale, market share, revenues, and profitability.

A more recent study by Deloitte & Touche and CEEM Information Services found that the 620 respondents in their survey of ISO 9000 registered companies reported annual cost savings ranging from $25,000 to $600,000, with an average annual savings of $179,000 (see Quality Systems Update [1993]). While this survey provides some evidence on the magnitude of cost savings, the savings are self-reported and depends on the perceptions and estimates of the respondents.

A common limitation of the above studies is that they do not test for the statistical significance of the improvements in performance. Additional weaknesses include the survey nature of the data and no controls for potential industry and/or economy wide influences.

Easton and Jarrell [1994] examine the impact of TQM practices on financial performance for a sample of 108 firms by comparing the actual performance with a benchmark performance measure of how the firm may have performed had they not adopted TQM. Statistical analyses are conducted. The results provide evidence of improved financial performance associated with implementation of TQM programs. The results are uniformly stronger for the subsample of 44 firms which they identify as having more mature TQM programs.

Recently, two studies examine the relation between TQM and stock price performance. Hendricks and Singhal [1994] show that the stock market reacts positively to announcements of winning quality awards. Statistically significant mean abnormal returns on the day of the announcements averaged 0.64%. The reaction was stronger for smaller firms (mean abnormal returns averaged 1.21%), and for firms that won awards from independent organizations such as the Baldrige Award, Philip Crosby, etc. (mean abnormal returns averaged 1.49%). The evidence indicates that large firms experience negative stock price performance in the second year before winning quality awards, which is followed by a year of positive performance. Small firms experience a positive stock price performance in the second year before winning quality awards, but no negative performance before winning quality awards. Hendricks and Singhal [1994] also document a statistically significant decrease in the systematic risk of the firm after the quality award announcement.

The other study is by Heller [1994], who reports that a portfolio of 150 TQM firms had a statistically significant abnormal change of 4.95% in stock prices over the years 1989-1992. Heller [1994] identifies the TQM firms by applying criteria developed from the Baldrige Award and other popular sources of information on TQM. Financial performance is not considered in identifying TQM firms.

A limitation of the studies by Hendricks and Singhal [1994] and Heller [1994] is that they only capture the partial impact of implementing effective TQM programs. Furthermore, these studies do not relate the increase in stock prices to improvements in operating performance. The increase in stock prices could also be due to capital market inefficiencies.

Other than the studies discussed above, we have been unable to identify any other rigorous empirical evidence on the issue of whether implementing effective TQM programs improves financial performance. Moreover, it appears to us that the existing evidence is not convincing one way or the other on this issue.

3. Concepts, hypotheses, and performance variables

3.1 Concepts

The concepts on how effective TQM programs will improve operating performance can be broadly classified into three areas: (1) costs of quality, (2) total customer satisfaction, and (3) organizational innovation. We briefly discuss these concepts to motivate the hypotheses examined in this study.

Cost of quality: This concept was developed by Juran [1951] in the early part of the TQM movement when the focus was on using quality control tools to improve the conformance dimension of quality. There are two competing theories on how improving the conformance level affects costs. Juran and Gryna [1980] develop the notion of an optimal conformance level by trading off the appraisal and prevention costs with the internal and external failure costs. They argue that the optimal conformance level implies a strictly positive proportion of defectives. On the other hand, Deming [1982] and Crosby [1979, 1984] prescribe that the optimal conformance level is zero defects. This prescription is based on the belief that producing higher conformance quality products is always less costly than producing low conformance quality products. This has led to the famous claim by Crosby that quality is “free”. Fine [1986] shows that when quality-based learning affects quality control costs, firms have incentives to push towards zero defects. Empirical evidence supports the views of Deming and Crosby. Garvin's [1983] study of the room air conditioning industry, and Abernathy, Clark, and Kantrow's [1981] study of the automobile industry, show that manufacturers with higher conformance quality have lower costs.

Improving conformance quality can also affect revenues. If the performance level of similar products offered by different firms is stable and prices are similar across different firms, a product with a higher conformance level has a better chance of gaining market share than a product with a lower conformance level. If customers perceive improvements in conformance quality, then they may be willing to pay higher prices. This could enable a firm to increase its revenues while maintaining its market share or vice-versa. In summary, the cost of quality concept suggests that improving conformance levels should increase profits.

Total customer satisfaction: Totally satisfying customer needs and expectations is a key element of any effective TQM program (see Garvin [1988], Schonberger [1990], and Kolesar [1993], among others). This point is also reinforced by most award givers as they assign a significant weight to how the award applicant has performed on customer satisfaction. Customer satisfaction is defined more broadly then simply providing a high quality product. It includes developing systems to determine customer expectations, establishing communications links and long-term relationships with customers, responding to customer needs in a timely manner, being committed to customers, developing customer satisfaction indicators, and taking actions to improve on these indicators. Higher customer satisfaction should generally lead to higher customer retention rate, increased market share, and higher profitability.

Organizational innovation: Many experts view TQM as a new organizational technology which enables the organization to utilize its human and other physical assets more productively. Key elements of this technology include training the work force in non-traditional approaches to problem solving, involving employees in decision making, delegating decision making and responsibility further down in the organization, teamwork, inter-functional problem solving effort, and changing the way employees are evaluated. For example, Wruck and Jensen [1994] argue that TQM is an efficiency improving organizational technology because 1) it encourages the use of the scientific method in every-day decision making at all levels of the organization; 2) it encourages the creation and utilization of specific knowledge, by transferring decision making rights to those agents who have the specific knowledge, and (3) it changes the performance measurement, reward, and punishment systems.

3.2 Hypotheses and performance variables

Based on the above discussion, our first hypothesis is that implementing an effective TQM program will improve the profitability of the firm. Our measure of profitability is operating income before depreciation which equals net sales less cost of goods sold and selling and administrative expenses before depreciation, depletion, and amortization are deducted. This measures the cash generated from operations before depreciation, interest, and taxes. It is therefore unaffected by the method of depreciating assets, the capital structure of the firm (debt/equity ratio), or the gains or losses from the sale of assets. Conceptually, it makes sense to focus on operating income before depreciation because it measures the economic value (cash flows). Note that the changes in operating income do not control for acquisitions and divestitures, which may result in mis-estimating the real change. To partially control for acquisitions and divestitures, other income based measures considered are annual operating income divided by year-end assets, by annual sales, and by year-end number of employees. Dividing by year-end assets measures return on assets, and dividing by annual sales measures operating margin.

Our second hypothesis is that implementing an effective TQM program will increase revenues. A commonly cited benefit of TQM programs is that it can lead to higher customer satisfaction which in turn could lead to higher sales (customers are willing to pay more or buy more). Net sales is the primary revenue measure used in our study. Other measures considered are asset turnover which is annual sales divided by year-end assets, and sales per employee which is annual sales divided by yearend number of employees.

Our third hypothesis is that implementing an effective TQM program will reduce costs. Reducing defects and rework, process improvements, and eliminating waste are key elements for implementing an effective TQM program. Progress in these aspects of TQM is expected to reduce costs. The primary cost measure used is the sum of annual cost of goods sold and selling, general and administrative expenses divided by annual sales (that is cost per dollar of sales).

We also explore the effect of implementing an effective TQM program on inventory, capital expenditures, number of employees, and total assets. There is an extensive literature on how quality management practices affect inventories. For example, by keeping the production process in control, work can flow faster through the system thereby reducing work-in-process (Schonberger [1982, 1986], and Monden [1983]). Similarly, if the quality of incoming material and outgoing products is improved, then the need for buffer inventories of raw material and finished goods decreases. Others have argued that there is a close linkage between Just-in-Time Manufacturing and Total Quality Management (see, for example, Sakakibara et. al. [1992]). Thus, one would expect that implementing an effective TQM program could reduce inventories. The primary measure used is inventory turnover defined as annual cost of goods sold divided by the average of the beginning and ending inventory in that year.

No specific hypotheses are offered for the direction of change in capital expenditure, number of employees, and total assets as arguments can be made for changes in either direction. For example, some believe that TQM programs require investment in people and capital, resulting in an increase in employment, capital expenditure, and total assets (see, for example, Port [1992], and Greising [1994]). Others believe that TQM programs increase the effective productive capacity of the firm because of process improvements and reductions in defects, rework, and waste, among other things. These improvements could result in a decrease in employment, capital expenditure, and total assets.

4. Sample selection and methodology

4.1 Sample selection

Two ways are used to identify firms that have won quality awards. First, by searching on-line databases using “Quality” and “Award” as key search words. Many award givers and/or award winners make public announcements about quality awards. Previous research indicates that such public announcements are typically available from two publication sources: PR Newswire and Business Wire (see Hendricks and Singhal [1994]). We did a comprehensive search of these two newswires which are archived in the Dow Jones News Service.

The second way used was to directly contact a number of award givers for names of their quality award winners. Table 1 gives the names of award givers who have provided information. The choice of award givers was influenced by information obtained from articles in trade publication on quality awards, and our discussion with a few TQM experts.

Table 1: Names of quality award givers that have provided names of their quality award recipients

Auto Alliance International Inc. (Part of Mazda Motor Manufacturing)

Chrysler Corp.

Consolidated Rail

Eastman Kodak Co.

Ford Motor Co.

General Motors Corp.

GTE Corp.

Honda of America Manufacturing Inc.

International Business Machines

Maryland Center for Quality and Productivity (The US Senate Productivity Award)

Minnesota Council for Quality

Minnesota Mining and Manufacturing

National Aeronautical and Space Authority

National Association of Manufacturers and Utah State University (The Shingo Prize)

National Institute of Standards and Technology (The Baldrige Award)

New United Motor Manufacturing Inc. (NUMMI)

Nissan Motor Manufacturing Corp. U.S.A

North Carolina Quality leadership Foundation

Philip Crosby (The Fanatics Award)

Sematech

Texas Instrument Co.

Toyota Motor Manufacturing U.S.A Inc.

Xerox Corp.

Our data base consists of nearly 4400 records of award receiving organizations paired with award giving organizations. Thus, a firm winning an award from three different award givers has 3 records. Nearly 100 different award givers are represented in our sample. The awards given by the award givers listed in Table 1 account for nearly 95% of the records in our database. Of the 4400 records, 1665 records (596 distinct firms) are of firms which have some information on the Compustat Annual Industrial File; 1247 records (1025 distinct firms) are of firms that are private; 438 records (205 distinct firms) are of foreign firms; and 1038 records (972 distinct firms) had no reference in information sources such as the Million Dollar Directory, Ward's Directory of Businesses, Standard and Poor's Directory of Corporations, Directory of Corporate Affiliations, Who Owns Who, and the Directory of Obsolete Securities. Probably these firms are too small to merit a reference in these publications or they may have been acquired by some other firms or even gone out of business. However, none of these firms were ever publicly traded because we have checked these names against the Directory of Obsolete Securities.

To be included in the test sample, a quality award winning firm must satisfy two criteria. First, the firm must be on the Compustat Annual Industrial File. And second, the firm must have continuous data for a minimum 6 fiscal years, beginning from six years before through one year before the year of winning the first quality award. The second criterion is imposed to reduce noise in our analyses. For example, suppose that implementing TQM requires significant investment in training and new equipment at the start of implementation, and that these investments pay off in subsequent years. In this case one would expect operating performance to worsen in the years around the start of implementation and then subsequently improve. Now consider a firm that was private during the years when it implemented the TQM program, and then subsequently went public. By keeping such a firm in our sample, we would be considering those years when the firm's operating performance is likely to be on the upswing. The reverse would be true if the firm was publicly traded during the implementation stage and not publicly traded subsequently. In either case, we would be introducing noise in our analyses.

There were 596 firms that met the first criterion, and 463 firms which met both criteria. For each of these 463 firms, we identified the year when the firm won its first quality award. Table 2 gives the distribution of the year of winning the first quality award for the 463 firms. Nearly 25% of the sample firms won their first quality award in 1989. About 32% of the sample firms won their first quality award during 83-88 whereas 43% won their first award during 90-93.

Table 2: The distribution of the calendar year when the 463 firms in the sample won their first quality award.

Table 3 gives summary statistics for the full sample of the 463 firms based on the most recent fiscal year completed before the year of winning their first quality award. The information was collected from the Compustat Annual File. The average observation in the data set represents a firm with net income of $189.5 million on sales of $4.57 billion with a market value of $2.69 billion. The smallest firm had sales of $7.7 million, while the largest firm had sales of over $127 billion. Net income ranged from a low of -$1.67 billion to a high of $5.26 billion. The number of employees ranged from a low of 82 to a high of 877,000. And the debt ratio (defined as the total debt divided by the sum of total debt and the market value of equity) similarly covered a wide range (from 2.1% to over 98%). In all, the sample has firms from 180 distinct four-digit SIC codes, and 42 distinct two-digit SIC Codes.

Table 3: Selected descriptive statistics for the sample of 463 firms for the most recent fiscal year ending prior to the winning of quality award

4.2 Choice of time period for analyses

To estimate the net benefits of implementing an effective TQM program, operating performance must be examined both before and after the effective implementation of TQM. The reason for examining performance before the effective implementation of TQM is that it is during this period that firms may incur significant one-time or ongoing direct and indirect costs in implementing TQM. The reason for examining the performance after the effective implementation of TQM is that it is during this period that the organization should realize the benefits, if any, of these programs. Furthermore, because implementing TQM requires time and patience, a suitably long time horizon must be considered to establish the relation between TQM and operating performance.

We examine the operating performance of our sample of quality award winners over a ten-year period. Ideally, we would have liked to anchor the beginning point of our estimation period on the year when firms initiate the implementation of TQM programs. Unfortunately, this information is not available as firms generally do not announce that they have started a TQM program. Furthermore, the existing literature does not provide much theoretical or empirical guidance on this issue. Some have suggested that it takes about 5 years to effectively implement TQM (King [1992], Hockman [1992], and and Easton and Jarrell [1994]). The U.S. General Accounting Office's (1991) study of 22 finalists and winners of the Baldrige award indicated that it took an average of about 2.5 years (the range was 1 to 5 years) to realize some initial benefits from TQM programs. Given this, we chose to start the estimation period six years before the year of winning the first quality award.

For most of the firms in our sample, the 1993 fiscal year is the last year for which data is available on the Compustat Industrial File (firms with year ends prior to June of 1994). The distribution of the year of winning the first quality award (see table 2) indicated that nearly 40% of the sample firms (firms that won their first quality award after 1989) would not have accounting information on the fourth year after the year of winning the first quality award. Given this, we chose to end the estimation period three years after the year of winning the first quality award.

To pool observations across time, for each firm in our sample, fiscal years are translated to event years using the following conventions. The fiscal year when the firm won its first quality award is denoted as year 0 in event year. The next fiscal year is denoted year +1 in event year, and the year after that as +2 and so on. The fiscal year before the year when the firm won its first quality award is denoted as year -1, and the year before this as year -2 and so on.

4.3 Creating matched-pair samples of comparison firms.

To provide a benchmark for the performance of our sample of quality award winning firms, and to control for potential industry and/or economy wide influences, we created three different control samples. We assume that firms in the same industry and of similar size are subject to similar economic and competitive factors. In the first control sample, for each quality award winning firm we chose a control firm that (1) has the same country of incorporation, (2) has accounting data available over at least the same time period as the award winning firms (that is over the years beginning six years before through three years following the year of winning the first quality award), (3) firms with fiscal year end in January to May (June to December) matched with firms whose fiscal year ends in January to May (June to December), (4) has at least the same two-digit SIC industry Code, and (5) is closest in size as measured by the book value of assets at the fiscal year-end before the winning of the quality award, with the constraint that the ratio of the book value of assets of the control and award winning firm is always less than a factor of 3.

We were able to find a control firm that met the above conditions for 335 (72%) of the 463 firms in our test sample. The median book value of assets for these 335 firms is $380 million, and insignificantly different from $418 million, the median book value of assets of the control firms. The percent of matches with the same two, three, and four-digit SIC industry codes are 53%, 21% and 26%, respectively. Fifty percent (50%) of the control firms have the same fiscal year-end, and 84% have fiscal year-end within three months of fiscal year-end of the award winning firms. We feel that this control sample gives the closest matches; however, it necessitated excluding nearly 128 firms (28% of the total sample).

After analyzing the excluded firms, it became obvious that very large firms were hard to match and constituted a large percentage of the excluded firms. To prevent any size bias, a second control sample was generated which augments the 335 matches obtained from the first control sample by attempting to find matches for the 128 unmatched award winning firms. This was done by relaxing the conditions on the fiscal year-end matching, and allowing a single-digit industry matching. Matching on single-digit SIC codes is more appropriate for large firms because they are likely to be more diversified than small firms. These new conditions were applied to only the 128 firms that we could not match in the first control sample.

We were able to match an additional 59 firms, bringing the total matches to 394 firms (or 85%) out of 463 firms in our test sample. As expected, the matching is good in terms of size. The median book value of assets for these 394 firms is $577 million, and insignificantly different from $588 million, the median book value of assets of the control firms. The percent of matches with the same one, two, three and four-digit SIC industry codes are 14%, 46%, 18%, and 22%, respectively. 54 of the 59 new matches are at the one-digit SIC industry code. 50% of the control firms have the same fiscal year-end, and 81% have fiscal year-end within three months of fiscal year-end of the award winning firms.

A third control sample was generated without any conditions on the fiscal year-end matching, and without the requirement that the control be within a factor of 3 of the size of the award winning firm. Specifically, when all the other conditions are met, we select as the control the one that is simply the closest in size. We could find a match for 431 (93%) of the 463 award winners. The only firms that are not matched are firms incorporated outside of the United States. The award winners and controls are mismatched on size. The median book value of assets for the 431 award winning firms is $811 million, and significantly higher than $330 million, the median book value of assets for the control firms (the Wilcoxon signed-rank test had a z -value of 5.99). The industry matching is very similar to the first control sample, whereas the fiscal year-end matching is very similar to the second control sample.

We feel that the first control sample gives the best overall matches, but with potential biases introduced by systematic differences in unmatched firms (very large firms not matched). Compared to the first control sample, the second control sample reduces these potential biases by adding another 59 firms by relaxing the industry matching while ensuring a good match on size. The third control sample matches all possible firms and does as well as the first control sample in terms of industry matching, but does poorly in terms of matching on size when compared to the first and second samples. Since, each control sample had its own strengths and weaknesses, we ran our analyses using all three control samples. Since, the results are similar across the three control samples, we only report the results obtained from the second control sample.

5.0 Empirical results

It is reasonable to expect that if a firm has an opportunity to apply for an award, it is likely to apply soon after it has implemented an effective TQM program. Furthermore, it can typically take the award giving organizations about 6 to 8 months to evaluate and certify the effectiveness of the program. Given this, we use year -1 as our best estimate of when the TQM programs of the award winning firms in our sample began to become effective. Thus, at the very least, we can claim that the performance after year -1 is indicative of what could be expected after implementing an effective TQM program.

We report two sets of results. The first set reports the difference in the percent change in the performance between the test and control samples on an annual basis over 10 years. The difference is computed by subtracting from the percent change in performance of each test firm, the percent change in performance of its control firm. These results identify time periods when operating performance significantly improves or deteriorates.

The second set of results reports the difference in percent change in performance between the test and control samples over longer time intervals with different starting and ending points. There are two reasons for this. First, the pattern of changes in performance could vary across firms. For example, there could be a firm that had significant deterioration in performance in year -3 because of say a major investment in training as part of the TQM program, whereas another firm could have had positive performance in year -3 because it had already started implementing its TQM program in year - 5. In any particular year, such negative and positive changes in performance could cancel each other out, and the net outcome could be a statistically insignificant change in performance. Second, the philosophy of continuous improvement is a key element of TQM programs. The thrust of this philosophy is on making small and incremental improvements on a regular basis, with the notion that small improvements add up to a significant improvement over longer time periods. If this is indeed the case, the changes measured on an annual basis may be small and statistically insignificant, yet could be significant over longer intervals.

More specifically, we report changes from years -6 to -1 (a five-year period before the evidence of effective TQM program), -4 to -1 (a three-year period before the evidence of effective TQM program), -1 to +1 (a two-year period after the evidence of effective TQM program), -1 to + 3 (a four-year period after the evidence of effective TQM program), and -6 to + 3 (the complete ten-year period). Recall that year -1 is our best estimate of when the TQM programs of the award winning firms in our sample began to become effective.

The mean and median results for the various performance variables are reported. To control for outliers, which can influence the mean values, all results are reported after symmetrically trimming the data at the 2.5% level in each tail. The results (not reported here) are similar using 5.0% trimming in each tail, and winsorizing at the 2.5% or 5.0% in each tail. Non-parametric results are basically the same with and without trimming and winsorizing. The Student t (Wilcoxon signed-sign rank) statistic is used to test whether the mean (median) change in the performance variable is significantly different from zero. All references to the significance levels of the results are based on one-tailed tests.

5.1. Differences in performance between the quality award winning firms and control firms.

5.1.1. Operating income based measures

Tables 4 through 11 report changes in various performance measures for the test sample after adjusting for the performance of the control. Table 4 reports the annual percent changes in operating income based measures. The mean and median changes in operating income of the test sample are higher than the control sample in seven out of the 10 years. The mean changes are positive and statistically significant in the following three years: -7 to -6 (15.87%, p-value = 0.001), -4 to -3 (14.38%, p-value = 0.002), and -1 to 0 (7.82%, p-value = 0.012). The median changes in these three years are also positive, with one year significant at the 1% level, and the other two years significant at the 6% level or better. Note that not all of the statistically significant results are positive. The mean changes are negative and statistically significant in the following two years: -5 to -4 (-7.97%, p-value = 0.014), and +1 to +2 (-7.45%, p-value = 0.032). The median changes in these two years are also negative, with one year weakly significant at the 10% level. Overall there are more years with statistically significant positive changes than negative changes. Furthermore, the magnitude of changes in the positive years are generally higher than the magnitude of changes in the negative years. The results for the ratio of operating income to assets, operating margin, and operating income per employee are similar to those for operating income.

Table 4: Mean and median percentage changes in operating income, and in the ratios of operating income to assets, to sales, and to number of employ ees for the sample of firms that won quality awards. Mean and median changes are reported on an annual basis after adjusting for the performance of the controls. The t - statistics for the mean, and the Wilcoxon z - value for the median are in parentheses.

Table 5 reports changes in operating income related measures over longer intervals. Over the ten-year period from -6 to + 3, the mean (median) control-adjusted change in operating income is nearly 79% (30%). The mean and median changes in operating income are highly significant with tvalue of 3.47 and z-value of 3.00, respectively. The results are similar for the mean changes in the ratio of operating income to assets, and operating margins. The mean change in operating income per employee is positive, but not statistically significant.

Table 5 Mean a nd median percentage changes in operating income, and in the ratios of operating income to assets, to sales, and to number of employees for the sample of firms that won quality awards. Mean and median changes are reported over varying time periods after a djusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

The results indicate that improvements in operating income start just before the winning of the quality award. For example, from years -6 to -1, and -4 to -1 (the years before the winning of quality awards) the changes in operating income measures are positive, but not statistically significant. However, from year -1 onwards operating income starts to improve. The changes in operating income measures from years -1 to +1, and from -1 to +3 are positive and significant. In particular, the results from years -1 to +3 are highly significant. In general, the mean changes are stronger and more significant than the median changes, but the results are similar.

Overall, the results provide strong evidence that firms that have won quality awards outperform a control sample on operating income based measures. As we have argued earlier, firms are likely to win quality awards if they have an effective TQM program in place. The evidence supports our hypothesis that implementing an effective TQM program improves the operating income performance of the firm.

5.1.2 Sales based measures.

Table 6 reports the annual percent changes in sales based measures. The mean and (median) change in sales of the test sample are higher than the control sample in eight (six) out of the ten years. The mean changes are positive and statistically significant in the following two years: 0 to +1 (1.89%, pvalue = 0.043), and +2 to +3 (4.07%, p-value = 0.001). The median change from years 0 to +1 is negative and insignificantly different from zero, and from years +2 to +3 it is positive and significantly different from zero at the 1% level. The results from years +2 to +3 holds for sales per employee. For sales per employee there is an additional significant negative result from years +1 to +2. For the ratio of sales to assets, there is a significant positive result from years -3 to -2, and a significant negative result from years -1 to 0. Overall, these results do not provide very strong evidence that firms in the test sample outperform the firms in the control sample on sales based measures.

Table 6: Mean and median percentage changes in sales, in the ratio of sales to assets, and in the ratio of sales to number of employ ees for the sample of firms that won quality awards. Mean and median changes are reported on an annual basis after adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

The evidence on control-adjusted changes in sales measures is stronger when change is measured over longer intervals. Table 7 shows that over the ten-year period from years -6 to +3, the mean (median) change in sales of the test sample is nearly 43% (18%) higher than the control. Both the mean and median change in sales are highly significant (at the 2% level or better). The results also suggest that firms in the test sample start outperforming the firms in the control sample just before the winning of the quality award. For example, from years -6 to -1, and -4 to -1 (the years before the winning of quality awards) the changes in sales are positive, but statistically insignificant. However, the changes in sales from years -1 to +1, and from -1 to +3 are positive and statistically significant. In particular, the mean change in sales from year -1 to +3 is 14.67 % with a t-value of 3.77 (the median change is 8.45% with a z-value of 3.06).

Table 7: Mean and median percentage changes in sales, in the ratio of sales to assets, and in the ratio of sales to number of employees for the sample of firms that won quality awards. Mean and median changes are reported for varying time periods after adjusting for t he performance of the controls. The t - statistics for the mean, and the Wilcoxon z - value for the median are in parentheses.

The mean change in the ratio of sales to assets from years -6 to +3, is positive and significant at the 5% level (the median is positive and significant at the 7% level). From -6 to -1, and -4 to -1, the median change in the ratio of sales to assets is positive and significant at the 5% level (the means are also positive and weakly significant at the 7% level). However, the change in sales per employee is insignificantly different from zero for all the time periods shown in Table 7.

Overall, the results provide reasonably strong evidence that firms that have won quality awards outperform a control sample in terms of growth in sales. The evidence supports our hypothesis that implementing effective TQM programs improves sales. This is consistent with some earlier studies that have shown that firms with better quality achieve higher sales (see, for example, Buzzell and Wiersema [1981], Craig and Douglas [1982] and Phillips, Chang, and Buzzell [1983], among others).

5.1.3 Cost based and other measures.

Table 8 reports the annual percent changes in cost per dollar of sales. The mean (median) changes in cost per dollar of sales of the test sample are lower than the control sample in seven (four) out of the ten years. The mean percent changes are significantly negative in three out of the ten years. These are from years -7 to -6, -4 to -3, and -1 to 0. The median changes in these years are negative and statistically significant at the 5% level in two out of these three years. None of the annual changes show a statistically significant increase in the cost per dollar of sales.

Table 8: Mean and median percentage changes in the ratio of total cost to sales, the ratio of capital expenditure to assets, and inventory turno ver for the sample of firms that won quality awards. Mean and median changes are reported on an annual basis after adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

Table 8 also indicates that the ratio of capital expenditure to assets decreased significantly from years -4 to -3 (the mean change is -7.32, p-value = 0.01, and the median change is -3.34%, p-value = 0.04). None of the other annual changes in capital expenditure are statistically significant. In addition, relative to the control, the inventory turnover performance of the firms in the test sample is significantly worse from years -4 to -3, and +1 to +2, and significantly better from years -1 to 0.

Table 9 reports the changes in these measures over longer intervals. From years -6 to +3, the mean change in cost per dollar of sales of the test sample is -0.95%, (insignificantly different from zero) when compared to the control. The mean change from years -1 to +3 is about -1%, and weakly significant at the 7% level. None of the mean changes over other intervals, and none of the median changes are significantly different from zero.

Table 9: Mean a nd median percentage changes in the ratio of total cost to sales, the ratio of capital expenditure to assets, and inventory turnover for the sample of firms that won quality awards. Mean and median changes are reported over varying time periods after ad justing for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

Table 9 also shows that from years -6 to +3, the median change in the ratio of capital expenditure to assets is 8.04%, significant at the 5% level (the mean change is 2.44%). Most of this change can be attributed to the increased rate of capital expenditure from years -6 to -1 (the median change in the ratio of capital expenditure to assets is 6.07%, significant at the 7% level). This could indicate that implementing TQM programs require that firms increase their capital expenditure, perhaps on better process control systems or better equipment. Finally, the evidence in table 9 suggests no statistically significant changes in inventory turnover.

Overall the evidence weakly suggests that firms in the test sample are more successful in controlling costs when compared to the firms in the control sample, and appear to increase their rate of capital expenditure. The inventory turnover performance of the test sample is no different from that of the control sample.

Table 10 and 11 report the percent changes in employment and total assets. The results indicate that most of the changes in employment and assets occur from year -1 onwards. The controladjusted mean percent change in employment is significantly positive from year 0 to +1. Similarly, the percent change in assets is significantly positive from years -1 to 0, and 0 to +1. Over the years -6 to +3, the median change in employment is 8.83%, whereas the median change from years -1 to +3 is 3.49%. The median changes in total assets over the same time periods are 18.26% and 6.89%, respectively. Overall, the firms in our test sample had higher growth in employment and assets when compared to the firms in the control sample.

Table 10: Mean and median percentage changes in the number of employees and total assets for the sample of firms that won qual ity awards. Mean and median changes are reported on an annual basis after adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

Table 11: Mean and median percentage changes in the number of employees and total assets for the sample of firms that won quality awards. Mean and median changes are reported over varying time periods after adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

5.2 Changes in performance of the quality award winning firms without adjusting for the performance of the control firms.

The control-adjusted results presented above probably underestimate the true change in operating performance from implementing effective TQM programs. There are two main reasons for this. First, the sample of quality awards winners is based on the information gathered from newswires and from award givers. It is conceivable that some of the firms in our control sample may have won quality awards from other award givers which may have not been announced. Or, they may have announced their awards in publications which were not accessed by our search. Secondly, while winning a quality award indicates an effective TQM program, not having won a quality award does not necessarily indicate an ineffective TQM program. For competitive reasons, some firms with effective TQM programs may choose not to apply for awards. Alternatively, firms may have been finalists for a competition type award where only a limited number of awards are typically given. This is in contrast to a qualification type of award where there is no limit on the number of winners as long as the winners meet the minimum criteria set by the award givers. There is no way to know which of the firms in our control sample may have an effective TQM program in place. Therefore, it may be instructive to examine the performance of the firms in our test sample without adjusting for the performance of the control, and get a sense of an upper bound on the impact of implementing effective TQM programs. Tables 12 and 13 present these results for selected performance measures.

Table 12 shows that firms in our test sample experienced large and highly significant positive changes (unadjusted for control) in operating income and net sales for nearly all ten years. Except from years +1 to +2, the mean annual percent change in operating income ranged from a low of 6% to a high of 26%, and for sales it ranged from 6% to 14%. The percent change in cost per dollar of sales decreased significantly from years -7 to -6 and -4 to -3, and increased significantly from years -5 to -4 and + 1 to +2. Similar conclusions are reached when changes for operating income and sales are measured over longer time periods (see table 13). From years -6 to -3 mean increases in operating income and sales are a highly significant 148.40% and 155.07%, respectively (median increases are 92.86% and 84.78%, respectively). The mean increase in cost per dollar of sales is 1.11% (significant at the 1% level), whereas the median increase is 0.73% (significant at the 5% level).

Table 12: Mean and median percentage changes in operating income, sales, and in the ratios of total costs to sales for the sample of firms that won quality awards. Mean and median changes are reported on an annual basis witho ut adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

Table 13: Mean and median percentage changes in operating income, sales, and in the ratios of total costs to sales for the sampl e of firms that won quality awards. Mean and median changes are reported for varying time periods without adjusting for the performance of the controls. The t-statistics for the mean, and the Wilcoxon z-value for the median are in parentheses.

We believe that the results in tables 12 and 13 probably overstate the true change in operating performance from implementing effective TQM programs. Given the distribution of the year when firms in our sample won their first quality award, our study uses data from 1977 to 1993, with most of the data from 1980 onwards. During this 17 year period, the United States economy was in an expansion state for nearly 15 years. There were three recessionary periods which lasted for a total of 29 months. Furthermore, the inflation rate in the late seventies and early eighties was high. For example, the annual increases in the consumer price index ranged from 7% to 13% from 1977 to 1981. Since 1982 the annual increases have ranged from 3 to 4%. These observations suggest that the growth in the economy and inflation could be partly driving the highly significant results reported in tables 12 and 13. Hence, these results must be interpreted with caution. It is also worth noting that most of the articles on TQM in the business press, and many studies report evidence on the economic impact of implementing a TQM program without adjusting for potential industry and/or economy wide factors. In our view, this may have created overly optimistic expectations of what can be achieved with TQM.

6. Discussion, summary, and future research directions

6.1 Discussion and summary

This study has explored the hypothesis that implementing effective TQM programs improves the operating performance of firms. The winning of a quality award is used as a proxy for the effective implementation of TQM programs. Overall the results provide strong evidence that firms that have won quality awards outperform a control sample on operating income based measures. Over a ten-year period starting six years before to three years after the year of winning the first quality award, the mean (median) change in the operating income for the test sample is 79% (30%) higher than that of the control sample. We also find that over the this same period, the mean changes in the ratios of operating income to assets, and to sales are also higher relative to the control by about 15%. The evidence supports the hypothesis that implementing effective TQM programs improves the operating performance of the firms. Furthermore, it suggests that criticisms about the systems of quality awards are unnecessary and misplaced. Organizations that give quality awards seem to give awards to those firms that subsequently show better financial performance.

The evidence indicates that there is not much improvement in operating income before the winning of the quality award. For example, over the years -6 to -1, and -4 to -1 (the years before the winning of the first quality award) the changes in operating income measures are positive but statistically insignificant. This suggests that implementing an effective TQM program may not necessarily result in poor performance during the implementation stage. This is important because managers often worry about the direct and indirect costs of implementing TQM programs. While these costs are real and often high, perhaps TQM programs provide at least some early benefits which outweigh the costs of implementation. However, note that the evidence indicates that from year -5 to -4, the award winning firms had a statistically significant decrease in operating income based measures.

Operating income measures generally improve from year -1 onwards. The changes in operating income measures from years -1 to +1, and from -1 to +3 are positive and significant. In particular, the results from years -1 to +3 are highly significant. This is consistent with the notion that once an effective TQM program is in place, then firms should experience improvement in performance. It could also be that firms are using quality awards to send a credible and verifiable signal about quality to the market place, and the improvement in performance is due to the favorable reaction by customers.

We find reasonably strong evidence that firms that have won quality awards do better on sales growth than the controls firms. Over the ten-year period the mean change in sales of the test sample is nearly 43% higher, and the median change nearly 18% higher than the control. Similar to the results on operating income, sales start improving just before the year of winning the first quality award.

We also find weak evidence that firms in the test sample have been more successful in controlling costs when compared to the firms in the control sample. From years -6 to +3, the change in cost per dollar of sales of the firms in the test sample is -0.95% when compared to the control. The mean change from years -1 to +3 is about -0.93%. The weak evidence on costs is somewhat surprising given the conventional wisdom that improving quality reduces costs. It is plausible that although firms have been able to reduce cost by adopting TQM, much of the cost reductions have been passed on to customers in the form of lower prices. A better way to examine the relation between quality and cost would be to use cost per unit instead of cost per dollar of sales. Unfortunately, data on cost per unit is not available from publicly released financial statements.

The ratio of capital expenditure to assets is higher for the test sample when compared to the control, and most of the increase can be attributed to the period before the winning of a quality award. This could suggest that firms may have made investments in process control and new equipment to implement TQM programs. The inventory turnover performance of the test sample is no different from that of the control sample. This issue needs to be explored further given the strong emphasis in the literature on the close relation between TQM and Just-in-time philosophies. Consistent with the growth in sales, we find evidence of growth in both employment and total assets.

6.2 Future Research Issues

There are number of avenues for future research. First, other methodologies and measures of performance could be used to estimate the economic impact of implementing effective TQM programs. For example, one could look at the long-term stock price performance of quality award winners, and test whether the results for the stock price performance are consistent with the results on operating performance based on accounting numbers. It would be interesting to see if firms that win awards from different award givers or win awards consistently do better over time than firms that do not win awards. Different award givers use different criteria for evaluating quality programs, and have different standards for giving awards. It would be interesting to see if these differences impact operating performance.

Second, there is considerable interest among managers and academicians in identifying '“best or effective” quality management practices. Practices can be labeled best or effective if they improve performance. Given the results of our study, it might be fruitful to focus on quality award winners to identify best practices. This could add to the existing empirical studies which describe and compare quality management practices across firms and countries (see, for example, Garvin [1984, 1986], and Ebrahimpour [1985, 1988]).

Third, future research could examine the characteristics of firms that have implemented effective TQM programs. These characteristics could include variables like size of the firm, how diversified is the firm, managerial and institutional ownership of the firm, capital structure of the firm, the structure of top managers' compensation, and the extent of competition faced by the firm, among others. Such studies could provide guidelines to the board of directors and top managers on how to create an environment where the organization as a whole is more likely to respond to TQM initiatives. A related issue is whether firms adopt TQM only when faced with a crisis situation.

Finally, recent research has documented that financial and organizational mechanisms such as management buyouts, leveraged buyouts, recapitalizations, and mergers have created value through significant improvements in operating performance. However, these studies do not provide much information on what operating actions led to the improvements in performance. It would be interesting to study if managers in these firms adopted elements of TQM, and if so what is unique about their implementation approach. This could also clarify the role of incentives and organization structure on the adoption of TQM.

References

Abernathy, W. J., K. B. Clark, and A. M. Kantrow, “The New Industrial Competition,” Harvard Business Review, 59, 4, (1981), 68-81.

Baker, G. P., and K. H. Wruck, “Organizational Changes and Value Creation in Leveraged Buyouts: The Case of the O. M. Scott & Sons Company,” Journal of Financial Economics, 25, (1989), 163-190.

Bartov, E., “Open-Market Stock Repurchases as Signals for Earnings and Risk Changes,” Journal of Accounting and Economics, 14, (1991), 275-294.

Benson, P. G., J. V. Saraph, and R. G. Schroeder, “The Effects of Organizational Context on Quality Management: An Empirical Investigation,” Management Science, 37, 9, (1991), 1107- 1124.

Bureau of Business Practices, Profiles of Malcolm Baldrige Award Winners, Needham Heights, MA, 1992.

Buzzell, R. D., and F. D. Wiersema, “Modeling Changes in Market Share: A Cross-sectional Analysis,” Strategic Management Journal, 2, (1981), 27-42.

Craig, C. S., and S. P. Douglas, “Strategic Factors Associated with Market and Financial Performance,” Quarterly Review of Economics and Business, Summer, (1982), 101-111.

Crosby, P. B., Quality is Free, McGraw-Hill, New York, NY, 1979.

Crosby, P. B., Quality Without Tears, McGraw-Hill, New York, NY, 1984.

Dale, B. G., and A. G. Duncalf, “Quality-related Decision Making: A study in Six British Companies,” International Journal of Operations and Production Management, 5, 1, (1985), 15-25.

Dann, L. Y., R. W. Masulis, and D. Mayers, “Repurchase Tender Offers and Earnings Information,” Journal of Accounting and Economics, 14, (1991), 217-251.

Deming, E. W, Quality, Productivity and Competitive Position, MIT Center for Advanced Engineering, Cambridge, MA, 1982.

Deming, E. W, Out of Crisis, MIT Center for Advanced Engineering, Cambridge, MA, 1986.

Easton, G. S., and S. L. Jarrell, “The Effects of Total Quality Management on Corporate Performance: An Empirical Investigation,” (1994), Working paper, School of Business, Indiana University.

Ebrahimpour, M., “An Empirical Study of American and Japanese Approaches to Quality Management in the United States,” International Journal of Quality and Reliability, 5, (1988), 5-24.

Ebrahimpour, M., “An Examination of Quality Management in Japan: Implications for Management in United States,” Journal of Operations Management, 5, 4, (1985), 419-431.

Ernst & Young and American Quality Foundation, International Quality Study: The Definitive Study of the Best International Quality Management Practices; Top-Line Findings, Ernst & Young, Cleveland, OH, 1992.

Fine, C. H, “Quality Improvements and Learning in Productive Systems,” Management Science, 32, 10, (1986), 1301-1315.

Fitzerald, C., and T. Erdmann, American Automotive Industry Action Group, Actionline, October 1992.

Flynn, B. B., R. G. Schroeder, and S. Sakakibara, “The Impact of Quality Management Practices on Performance and Competitive Advantage,” Working Paper, (1993), Carlson School of Management, University of Minnesota, Minneapolis.

Foster, G., Financial Statement Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1986.

Fuchsberg, G, “Total Quality is Termed Only Partial Success,” Wall Street Journal, October 1, (1992a), B1.

Fuchsberg, G, “Quality Programs Show Shoddy Results,” Wall Street Journal, May 14, (1992b), B1.

Garvin, D. A., “How the Baldrige Award Really Works,” Harvard Business Review, 69, 6, (1991), 80-94.

Garvin, D. A., Managing Quality, Free Press, New York, 1988.

Garvin, D. A., “Quality Problems, Policies, and Attitudes in United States and Japan: An Exploratory Study,” Academy of Management Journal, 29, (1986), 653-673.

Garvin, D. A., “Japanese Quality Management,” Columbia Journal of World Business, 19, 3, (1984), 3-12.

Garvin, D. A., “Quality On the Line,” Harvard Business Review, 61, 4, (1983), 65-75.

Grant, R. M., R. Shani, and R. Krishnan, “TQM's Challenge to Management Theory and Practice,” Sloan Management Review, Winter, (1994), 25-35.

Greising, D., “Quality: How to Make it Pay,” Business Week, August 8, (1994), 54-59.

Haim, A., “Does Quality Work? A Review of Relevant Studies,” Conference Board, Report Number 1043, (1993), New York.

Healy, P. M., K. G. Palepu, and R. S. Ruback, “Does Corporate Performance Improve After Mergers?,” Journal of Financial Economics, 31, (1992), 135-175.

Heller, T., “The Superior Stock Market Performance of a TQM Portfolio,” The Center for Quality Management Journal, 3, 1, (Winter 1994), 23-32.

Hendricks, K. B., and V. R. Singhal, “Quality Awards and the Market Value of the Firm: An Empirical Investigation,” Forthcoming in Management Science, (1994),

Hite, G. L., and M. R. Vetsuypens, “Management Buyouts of Divisions and Shareholder Wealth,” Journal of Finance, 44, 4, (1989), 953-970.

Hockman, K. K., “Does the Baldrige Award Really Work,” Harvard Business Review, 70, 1, (1992), 137.

Jensen, M. C., and W. H. Meckling, “Specific and General Knowledge, and Organization Structure,” in Contract Economics, Lars Werin and Hans Wijkander, editors, Basil Blackwell Ltd., Oxford, United Kingdom, 1992.

John, K., L. H. P. Lang, and J. Netter, “The Voluntary Restructuring of Large Firms in Response to Performance Decline,' Journal of Finance, 47, 3, (1992), 891-917.

Johnson, T. H., “To Achieve Quality, You Must Think Quality,” Financial Executive, (May- June 1993), 9-11.

Juran, J. M., and F. M. Gryna, Quality Planning and Analysis, McGraw-Hill, New York, NY, 1980.

Juran, J. M., Quality Control Handbook, McGraw-Hill, New York, NY, 1951.

Kaplan, S., “The Effects of Management Buyouts on Operating Performance and Value,” Journal of Financial Economics, 24, (1989), 217-254.

Kelly, K., “Quality: Small and Midsize Companies Seize the Challenge - Not a Moment Too Soon,” Business Week, (November 30, 1992), 66-69.

King, R. “Using Total Quality Management to Improve Bottom-Line Results,' GOAL/QPC 9th Annual Conference, October 26, (1992), Boston, MA.

Kolesar, P., “Scientific Quality Management and Management Science,' in Logistics of Production and Inventory: Handbook of OR and MS, 4, S. C Graves et. al. eds.(1993).

Lichtenberg, F. R., and D. Siegel, “The Effect of Leveraged Buyouts on Productivity and Related Aspects of Firm Behavior,” Journal of Financial Economics, 27, (1990), 165-194.

Mathews, J., and P. Katel, “The Cost Of Quality: Faced With Hard Times, Business Sours on Total Quality Management,” Newsweek, September 7, (1992), 48-49.

Monden, Y., Toyota Production Systems: Practical Approach to Production Management, Institute of Industrial Engineers, Norcross, GA, 1983.

Muscarella, C. J., and M. R. Vetsuypens, “Efficiency and Organizational Structure: A Study of Reverse LBO's,” Journal of Finance, 45, (1990), 1389-1414.

Ofek, E., “Efficiency Gains in Unsuccessful Management Buyouts,' Journal of Finance, 49, 2, (1994), 637-654.

Peach, R. W, “Creating a Pattern of Excellence,” Target, 6, 4, (1990), 15-22.

Phillips, L. W., D. Chang, and R. D. Buzzell, “Product Quality, Cost position, and Business Performance: A Test of Some Key Hypotheses,” Journal of Marketing, 47, (Spring 1983), 6-43.

Port, O., “Quality: Small and Midsize Companies Seize the Challenge- Not a Moment Too Soon,” Business Week, (November 30, 1992), 66-72.

Quality Management Update, “IBM's Good News,” January-February, 1993.

Quality Systems Update, Special Report, CEEM Information Services, September, (1993), Fairfax, Virginia.

Rath and Strong, Fall 1991 Survey on Current Trends in Implementing Total Quality: Summary of Findings, Rath & Strong Management Consultants, Lexington, MA, 1992.

Reimann, C. W., and H. S. Hertz, “The Malcolm Baldrige National Quality Award and ISO 9000 Registration,” ASTM Standardization News, November, (1993), 42-53.

Sakakibara, S., B. B. Flynn, R. G. Schroeder, and W. T. Morris, “The Impact of Just-in-time Manufacturing and its Infrastructure on Manufacturing Performance,” Working Paper (1992), Carlson School of Management, University of Minnesota, Minneapolis.

Saraph, J. V., P. G. Benson, and R. G. Schroeder, “An Instrument for Measuring the Critical Factors of Quality Management,” Decision Sciences, 20, 4, (1989), 810-829.

Schonberger, R. J., Building a Chain of Customers, Free Press, New York, 1990.

Schonberger, R. J., World Class Manufacturing: The Lessons of Simplicity Applied, Free Press, New York, 1986.

Schonberger, R. J., Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity Free Press, New York, 1982.

Senge, P., “Quality Management: Current State of the Practice,” Keynote Speech at the American Society of Quality Control Annual Conference, 1993.

Smith, A , “Corporate Ownership Structure and Performance: The Case of Management Buyouts,” Journal of Financial Economics, 27, (1990), 143-164.

Szwergold, J., “Why Most Quality Efforts Fail,” Management Review, August, (1992), 5.

United States General Accounting Office, Management Practices, U.S. Companies Improve Performance Through Quality Efforts, 1991, (GAO/NSIAD-91-190), Washington DC.

Wruck, K. H., and M. C. Jensen, “Science, Specific Knowledge and Total Quality Management,” Journal of Accounting and Economics , 18, 3, (1994), 247-287.

Wruck, K. H., “Financial Policy, Internal Control, and Performance: Sealed Air Corporation's Leveraged Special Dividend,” Forthcoming in the Journal of Financial Economics , (1994), Working Paper, Harvard Business School, Boston, MA.