1. Introduction
The modern firm is a key player in market economies.Footnote 1 Public confidence that companies do not cheat, shirk or act otherwise opportunistically seems to be essential for the legitimacy and general acceptance of a market economy order (e.g. Di Tella and MacCulloch, Reference Di Tella and MacCulloch2009). Lack of confidence is associated with a skeptical view of a capitalist system and a positive attitude toward market regulation (Aghion et al., Reference Aghion, Algan, Cahuc and Shleifer2010; Dimitrova-Grajzl et al., Reference Dimitrova-Grajzl, Grajzl and Guse2012; Pinotti, Reference Pinotti2012; Pitlik and Kouba, Reference Pitlik and Kouba2015). Assuming that confidence hinges on rules, formal procedures, and practices that shape how organizations function, economic regulation can help to increase confidence in major companies as key market organizations (Anania and Nisticò, Reference Anania and Nisticò2004; Carlin et al., Reference Carlin, Dorobantu and Viswanathan2009).
Yet there is surprisingly little empirical work on the foundations of confidence in companies in general. Research based on individual-level data focuses predominantly on political and legal institutions and organizations. A central finding is that generalized social trustFootnote 2 is associated with higher confidence in parliament, public administration or the legal system, although the channels and direction of causality are not always clear (e.g. Newton et al., Reference Newton, Stolle, Zmerli and Uslaner2018; Rothstein and Stolle, Reference Rothstein, Stolle, Castiglione, van Deth and Wolleb2008; Sønderskov and Dinesen, Reference Sønderskov and Dinesen2016). The few studies that analyze confidence in the business sector or in companies either focus on individual-level determinants or refrain from investigating effects of economic regulation, a key policy variable (Adams et al., Reference Adams, Highhouse and Zickar2010; Chan et al., Reference Chan, Lam and Liu2011; Harms and Schwab, Reference Harms and Schwab2019; Kim, Reference Kim2012; Pirson et al., Reference Pirson, Martin and Parmar2019; Uslaner, Reference Uslaner2010). We attempt to narrow this research gap and investigate empirically the relationship between generalized social trust, the intensity and quality of business regulation, and confidence in major companies.
In section 2 we present our main hypotheses. We describe data and variables in section 3. In section 4 we outline our empirical methodology and in section 5 we present and discuss our results. Section 6 concludes the study.
2. Theoretical background and testable hypotheses
Two different views on the main sources of confidence are advanced in the literature (e.g. Camussi and Mancini, Reference Camussi and Mancini2019; Mishler and Rose, Reference Mishler and Rose2001): the “cultural” and “experiential” approaches.
Generalized social trust and confidence in major companies
“Cultural” theories for the formation of generalized social trust emphasize the importance of early socialization in families, kindergarten, and school (Uslaner, Reference Uslaner2008). Trustful individuals have a more optimistic view of the behavior of others and thus tend to be trustful in most circumstances (Newton et al., Reference Newton, Stolle, Zmerli and Uslaner2018). From this viewpoint an individual's confidence in institutions and organizations alike is an extension of their generalized trust of other people (Mishler and Rose, Reference Mishler and Rose2001), and thus mostly exogenous to the performance of the respective political and market institution or organization.
However, generalized trust is directed horizontally toward “equals,” while confidence in political organizations refers to a vertical relationship with “authorities” (Newton, Reference Newton, Russell and Klingemann2007). The impact of trust in unknown people on confidence in companies should be stronger if the relationship with the business sector is also perceived as horizontal, as it might spill over more easily. If, on the contrary, the relationship between companies and individuals, as customers, clients, or employees, is seen as “not among equals,” with companies ostensibly exercising real power, attitudes to major companies will be shaped by social trust to a lesser extent.
Whether individuals perceive companies as equals or as authorities is – to the best of our knowledge – unexplored. It is yet conceivable that small local firms (e.g. the neighborhood barber) fall in the first group, while major companies are more recognized as authorities.Footnote 3 Generalized trust should then be a stronger predictor for confidence in smaller firms compared to bigger companies as a consequence of a (gradually) different perception of the relationship as more horizontal (small business) rather than hierarchical (major companies).
Taken together, an individual's trust in unknown others is expected to be positively associated with confidence in major companies, especially if the relationship with major companies is perceived as horizontal. Based on these conceptual considerations we derive:
Hypothesis 1:
Social trust is positively related to confidence in major companies.
Regulation and confidence in major companies
“Experiential” theories stress that confidence depends crucially on organizational performance (Helliwell et al., Reference Helliwell, Wang and Xu2016). Bad performance is a potential source of lack of confidence, and good performance in the business sector enhances confidence in companies at large.Footnote 4 However, it is quite unclear how “collective” performance of the business sector is assessed. To analyze potential sources of confidence in business one needs to detect the obligations that business has to society (Baumol, Reference Baumol2016; Cohen, Reference Cohen, Harris, Moriart and Wicks2016), which can then be evaluated against a process- or a results-oriented benchmark.
The process-oriented view is based on the notion that people judge performance by perceptions of fair treatment and ethical behavior. Low confidence in political organizations is often associated with corruption, favoritism and administrative slack (e.g. Clausen et al., Reference Clausen, Kraay and Nyiri2011; Grönlund and Setälä, Reference Grönlund and Setälä2012). With respect to the business sector, a similar effect is likely to hold when bribing, fraud, greed, and unethical behavior of managers are pertinent (e.g. Glazer et al., Reference Glazer, Kanniainen and Poutvaara2010; Edelman, 2016). Regulatory rules that protect against opportunistic behavior, exploitation, and unlawful conduct may thereby contribute to confidence in firms (Anania and Nisticò, Reference Anania and Nisticò2004; Carlin et al., Reference Carlin, Dorobantu and Viswanathan2009).Footnote 5 Berggren and Jordahl (Reference Berggren and Jordahl2006) claim that regulatory rules and unbiased courts that establish and enforce private property rights in a fair and efficient manner promote expectations that people in general behave trustworthily.Footnote 6 Experimental studies also find that competition fosters trust as it gives stakeholders leeway as they decide with whom to interact (Huck et al., Reference Huck, Lünser and Tyran2012). Repeated interactions in a competitive and well-regulated market environment thus generate trust in market actors.
Regulatory interventions could however also increase suspicion of companies. People interpret regulation as a signal that “something is wrong” and actually lose confidence. Excessive over-regulation as well as under-regulation generate incentives for rent seeking and opportunities to companies for misconduct. Consequently, it is not the intensity dimension (i.e. the extent) of regulation per se that is key for confidence in major companies. Rather, the process with which rules are enforced is supposedly decisive.Footnote 7 If business regulation serves narrow interests it undermines confidence.Footnote 8 Based on this discussion we hypothesize:
Hypothesis 2:
Confidence in major companies is enhanced by impartial business regulation.
Macroeconomic performance and confidence in major companies
The performance of companies can also be evaluated by a results-oriented assessment of the business sector. For clients and customers, price and quality of provided goods and services are main indicators for a company's performance. For shareholders, company performance is probably best approximated by long-term profits and the development of share prices. For employees, salaries and working conditions matter most. Business scandals are probably important for all stakeholders (Bowler and Karp, Reference Bowler and Karp2004). All factors are however “particularized” and are hardly crucial in predicting universal public attitudes toward major companies.
Literature on vote popularity functions (Lewis-Beck and Stegmaier, Reference Lewis-Beck and Stegmaier2013; Nannestad and Paldam, Reference Nannestad and Paldam1994) suggests that unemployment rates, inflation, and GDP growth have significant influence on the popularity and electoral success of political office holders. Related papers indicate that these factors also impact on confidence in business. Stevenson and Wolfers (Reference Stevenson and Wolfers2011) report evidence that high unemployment in the US is connected to significant declines of confidence in “big business” and major companies, and with an albeit insignificant drop in the perceived honesty of business executives. Countries that experienced an upsurge of unemployment during the Great Recession observed a strong decline of confidence in government and in the finance sector. Employing US data over a period of 40 years, Kenworthy and Owens (Reference Kenworthy, Owens, Grusky, Western and Wimer2011) likewise report that attitudes toward business and finance somewhat sour during recessions. Hence, based on the results-oriented view on companies’ obligations to society we expect:
Hypothesis 3:
Confidence in major companies is enhanced by good macroeconomic performance.
3. Data and variables
Confidence in major companies and generalized social trust
The main source for individual-level data in our analysis is the integrated World Values Survey/European Values Study (WVS/EVS; version v2018-09-12). WVS/EVS is a global research project that collects data on personal values and beliefs, as well as how they change over time. Representative country samples have a minimum of 1,000 interviews per survey wave. Surveys are conducted face to face over several months during the survey year. For data availability reasons we use all survey waves from 1995 until 2014 and confine our analysis to democratic systems. The resulting pseudo-panel comprises more than 100,000 respondents from 55 countries all over the world. It should be noted that the timing of the surveys opens the possibility that economic shocks during the survey period have an impact on our analysis that is based on annual data. However, as we confine our analysis to democratic countries such incidences should be rather rare.
Our dependent variable Confidence in major companies is obtained from the question “I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it a great deal of confidence, quite a lot of confidence, not very much confidence or none at all?” One item mentioned is “major companies.” Asking respondents specifically about major companies implies that answers signal confidence in anonymous organizations. Asking about perceptions of companies in general might invite respondents to think about the local shop in the neighborhood, for example, which would comprise particularized trust rather than confidence. The WVS/EVS question thus helps to avoid mixing up two related but still distinct concepts.
Originally coded on a 4-point Likert-scale, we recode Confidence in major companies as a dummy variable with value 1 if the respondent's answer was “a great deal” (original scale 4) or “quite a lot” (original scale 3), and 0 otherwise, as we are not interested in fine-grained differences in the strength of confidence. Also, recoding aims to reduce measurement issues. While responses of 4 or 3 clearly show that a respondent has confidence in major companies, it is not evident what exactly the difference between, say, “great deal” and “quite a lot” measures. Uslaner (Reference Uslaner, Lyon, Moellering, Sanders and Hatzakis2015, p.77) stresses that fine-grained scales encourage “the tendency of people to respond … by ‘clumping’ their answers around the mean” and that “respondents don't discriminate well in the scores they assign.” For our application this means that people tend to choose values 2 or 3 too frequently. Despite these concerns, in a sensitivity analysis we model confidence in major companies on the original 4-point scale.
Social trust, also measured at the individual level, is a dummy variable that takes the value 1 if the interviewed person approvingly answers the value survey statement that “Most people can be trusted.” If the reply is “You can't be too careful,” the indicator is coded 0.
Quality and quantity of business regulation
The quality and the quantity dimensions of economic regulation are derived from components of the Economic Freedom of the World's (Gwartney et al., Reference Gwartney, Lawson, Hall and Murphy2018) area 5 index, which measures how government regulations restrict entry into markets and interfere with the freedom to engage in voluntary exchange.
We capture Impartiality of business regulation with the sub-component “extra payments/bribes/favoritism,” based on Global Competitiveness Report's questions regarding views of a representative sample of business leaders in their respective countries. Survey questions cover perceptions about undocumented extra payments connected with licensing, public contracts, or judicial decisions, illegal payments aimed at influencing government policies, laws, or regulations, and the extent that government officials show favoritism to well-connected firms and individuals. Thus this sub-index provides information on perceived corruption, clientelism, patronage, and special favors to particularized interests in the process of making and enforcing rules.
In a robustness check we approximate quality of regulation by the broader EFW index “Legal structure and security of property rights.” This is a compound indicator derived from various sources covering several aspects of procedural quality of the legal system, such as judicial independence, impartiality of the courts, protection of property rights, integrity, or effective enforcement of contracts. It is best understood as a norm regarding the exercise of public power, and is not necessarily linked with distinct content of policies. However, it may also capture an efficient criminal justice system that supports confidence in major companies, since people then know that cheating companies will be brought to justice and punished.Footnote 9
The Extent of business regulation, i.e. whether markets are regulated and by how much market competition is restricted, is judged by other EFW regulation area 5 sub-components measuring the bureaucracy costs of regulatory compliance and administrative requirements, bureaucratic inefficiency, as well as time and costs of starting a business.
We also include sub-indices for the extent of Credit market regulation and Labor market regulation in the analysis. This is done since respondents, when thinking about major companies, may have in mind big financial organizations or key employers. All EFW indices are coded on a 0–10 scale. Higher values indicate higher quality, fewer restrictions on competition, and free exchange on the markets.Footnote 10
Macroeconomic performance
We rely on two indicators of macroeconomic performance in the respective survey year: the growth rate of real GDP per capita (GDP growth) and the unemployment rate (Unemployment rate), both from the World Development Indicators (WDI) database.
Control variables
The EVS/WVS dataset contains standard micro-level control variables used in related literature. Age, relative income, and a respondent's employment status (employed, self-employed, unemployed, retired) capture the possible effects of direct experience with companies. Education, religiousness and religious denomination may be considered as channels of third-party accounts about major firms that potentially shape confidence. Gender captures possible differences in the moral judgments of males and females.
We include a variable indicating the individual political self-assessment of respondents, expecting left-oriented people to have less confidence in major companies. Left-wing political orientation is measured on a 10-point scale, with higher values implying a more left-leaning positioning. Chan et al. (Reference Chan, Lam and Liu2011) show that the degree of a respondent's inequality aversion matters for confidence in big companies in the USA. Therefore, we add a variable capturing a respondent's income inequality attitudes to our set of regressors (Income equalization). The relevant WVS/EVS question asks whether “Incomes should be made more equal” or “We need larger income differences as incentives.” Replies are coded on a 10-point Likert-scale, with higher values indicating stronger income equalization preferences. Lastly, we include the (logarithm of) GDP per capita to capture the influence of economic development. Data for GDP per capita are taken from the WDI database.
Table A1 in the appendix displays descriptive statistics for the variables considered in the analysis, and Table A2 contains pairwise correlation coefficients for the variables of main interest. Correlation coefficients are generally low. The exception is the correlation between Extent of business regulation and GDP per capita, with a value of 0.77. It seems that wealthier nations have less business regulation in force (e.g. Voigt et al., Reference Voigt, Gutmann and Feld2015).
4. Empirical methodology
We estimate variants of the following empirical model:
F signifies the standard normal cumulative distribution; i = individual, j = country, and t = year; γ t and α j are year- and country-fixed effects;Footnote 11 k and l signify the number of control variables included in $Z_{jt}^{}$
and $X_{ijt}^{}$
. $Z_{jt}^{}$
includes the two variables capturing the macroeconomic performance, (log of) GDP per capita and the extent of Credit market and Labor market regulations, and $X_{ijt}^{}$
contains a battery of individual-level covariates.
Given the hypotheses derived in section 2 and the discussion of the measurement of key variables in section 3, we expect the coefficients on Social trust, Impartiality of business regulation and GDP growth to be positively signed. The unemployment rate is expected to enter equation (1) with a negatively signed coefficient. As argued in section 2 the sign of the variables capturing the extent of regulation is ambiguous a priori. Specifically, if less business regulation is conducive to confidence we expect β 3 to be > 0.
Due to the nature of the endogenous variable we apply the probit estimator to equation (1). We show cluster-robust standard errors with clustering at the country level. We consider 55 democratic countries (cf. the Notes to Table A1 for the list of included countries) with a Freedom House/imputed Polity2 democracy score of 8 and over (Hadenius and Teorell, Reference Hadenius and Teorell2007) in the analysis. The focus on democracies leaves us with a relevant number of clusters for statistical inference and reduces concerns about parameter stability across democracies and non-democratic regimes.Footnote 12
5. Results
Baseline results
The first two columns of Table 1 contain our baseline results. Column (1) excludes variables capturing Credit and Labor market regulation, while column (2) includes the full set of regulation variables. Note that in Table 1 we show average partial effects (APEs), which allow for a substantive interpretation of regression results.Footnote 13
Table 1. Confidence in major companies: Probit Estimates
Notes: *p < 0.10, **p < 0.05, ***p < 0.01; Average Partial Effects (APEs) are shown; APEs average the partial effect across the distribution of the respective variable; cluster-robust standard errors in parenthesis; clustering on country level with 55 clusters; higher values of the regulation variables imply less strict regulations (Extent, Credit Market Regulation, Labor Market Regulation) and higher impartiality / quality of regulation (Impartiality and Impartiality of the Legal System).
Respondents who trust unknown others have a higher probability of having Confidence in major companies. The coefficient is stable across the two models estimated. The APEs on Social trust imply that persons who trust unknown others have a 4 percentage points higher probability of showing confidence in major companies. Thus, while Social trust has a positive statistical association with Confidence in major companies, its substantive impact is rather limited. This finding is consistent with the view that respondents perceive major companies as vertical rather than horizontal.
Impartiality of business regulation unequivocally increases confidence in major companies for a given Extent of business regulation (Column 1) and of Credit market regulation and Labor market regulation, respectively (Column 2). An increase in Impartiality by 1 point increases the probability of confidence in major companies between 3 and 4 percentage points. The lowest value of Impartiality in our sample is about 2.7 (Venezuela in 1996), the average value is about 6.7 and the maximum value is 9.9 (Iceland in 1999). Our results imply that if Venezuela improved the quality of business regulation to an average level, the probability of Confidence in major companies would increase by about 12 to 16 percentage points.
The positive coefficient of Extent of business regulation and of Credit market regulation implies that for a given quality of business regulation, easing the regulations increases confidence in major companies. This indicates that the “bad parts” of regulation are dominant and that rent-seeking aspects are more important than public interest facets. In contrast, the negative (albeit insignificant) coefficient on Labor market regulation signals that the weaker hiring & firing regulations, stipulations regarding minimum wages, or collective wage bargaining arrangements are, the less confidence respondents have. Restrictions on firms in their scope to structure their workforce are conducive to generating confidence in major companies.
Table 1 reveals that bad macroeconomic performance diminishes confidence in major companies. A decrease in Unemployment rate (GDP growth) by 1 percentage point increases (decreases) respondents’ likelihood of showing confidence in major companies by about 1 percentage point. The minimum Unemployment rate in our estimating sample is 2.2% and the average level is about 7.9%. Thus, if an average country reduced its unemployment rate to the lower benchmark, confidence in major companies would – ceteris paribus – increase by about 6 percentage points. The average GDP growth rate in our sample is 2% with a maximum value of 12%. If an average country observed a growth rate at the upper benchmark this would result in an increase in confidence of 10 percentage points.
Turning to individual-level variables confirms expectations that confidence is lower the more pronounced the left-wing political orientation of respondents is. This likely signals their general negative attitude toward capitalist systems, for which major companies are quasi-emblematic. Likewise, people with higher income inequality aversion show lower confidence in major companies.
More educated people express lower confidence. As asserted by Uslaner (Reference Uslaner2010) this may reflect the greater attention of educated people to news about scandals and unethical behavior of companies in general. Respondents with middle and high income show higher confidence than low-income respondents. This might represent the more positive direct experience of higher-income earners with major companies. Confidence is also lower in females.
Religious people have higher confidence in major companies than non-religious people. This is in line with Guiso et al. (Reference Guiso, Sapienza and Zingales2003), who find that religious people generally trust more. With respect to religious denomination we find that people with Roman Catholic, Protestant, Muslim, and Jewish faith express higher probabilities of confidence compared to respondents with other religious denominations. The 31–60 age group, the prime working group, shows lower confidence than older and younger cohorts. Self-employed and unemployed people show lower confidence in major companies than employed persons, indicating that both project their negative experiences in interactions with major companies onto their confidence in these organizations. Regarding self-employed persons this finding is consistent with the view that people who run small businesses see major companies as a threat. Lastly, our results imply that confidence in companies is lower in democracies with higher GDP per capita.
In column (3) of Table 1 we explore whether our variables of main interest still show a significant association with Confidence in major companies when we control for a respondent's confidence in political organizations. A person's confidence in major companies and their confidence in political organizations likely share common determinants (Roussey and Deffains, Reference Roussey and Deffains2012). Inclusion of a closely related confidence variable among predictor covariates makes it more difficult to a find significant association of other variables with Confidence in major companies. While we admit that including confidence in a political organization may raise concerns about reverse causality in the estimations, we also believe that finding evidence consistent with our hypotheses even in this case strengthens the credibility that our variables of main interest matter specifically for Confidence in major companies. We opt for confidence in parliament as additional control variable as it is likely a less partisan organization compared to confidence in government, which is possibly biased by political orientation of the respondent.
As expected, Confidence in parliament is highly positively correlated with Confidence in major companies (cf. column (3)). In addition, we see that the coefficient of Social trust drops in magnitude signifying the frequently encountered association of generalized trust and confidence in political organizations in empirical studies. However, even after controlling for confidence in the parliament, Social trust still shows the hypothesized statistical association with Confidence in major companies. Furthermore, Impartiality of business regulation keeps its positive association. The only major change in results is with respect to the unemployment rate, which lacks statistical significance in column (3).
In column (4) we substitute Impartiality of business regulation by Impartiality of the legal system. While still positively signed this variable falls short of statistical significance. We interpret this finding as indicating that it is specifically the unbiasedness of business regulation that matters in strengthening confidence in the business sector.
Endogeneity and robustness
Taken together, results presented in Table 1 provide evidence in favor of Hypotheses 1 to 3. Yet, there is always the apparent issue of causality when it comes to the relationship of social trust, confidence, and regulatory provisions. First, a lack of both confidence in companies and of generalized social trust may cause higher demand for government intervention. However, we want to stress that we investigate links between country-level regulation and individual-level confidence (a mixed micro/macro approach). It is rather unlikely that the (lack of) confidence of one particular person substantially influences policies at the national level. This fact reduces a potential reverse causality problem (also see Landier et al., Reference Landier, Thesmar and Thoenig2008).
Second, one might be concerned about a non-causal relationship between Confidence in major companies and Social trust, not least as both are values of the same respondent and deal with closely related concepts. We delve into this non-causality concern and instrument Social trust by the product of individual-level variable Tolerance and the time-invariant macro-variable Monarchy. The EVS/WVS variable Tolerance asks respondents about important child qualities: tolerance and respect for other people. Thus, this variable concerns perceptions of other people and Tolerance should be sufficiently (positively) correlated with Social trust. Monarchy captures whether respondents live in a monarchy. Empirical literature shows that citizens of monarchies have higher generalized social trust (Bjørnskov, Reference Bjørnskov2007). Monarchy is measured at the country level and is time-invariant. As we include country fixed effects in the analysis we cannot use Monarchy per se as instrument. Yet by interacting it with Tolerance we can identify its impact on Social trust.Footnote 14
Table 2 includes results from bivariate probit regressions where we instrument Social trust with Tolerance X Monarchy. Column (2) shows that our instrument is a strong predictor of Social trust (F-value > 25). The coefficient of the instrument also has the expected positive sign. From column (1) we see that Social trust virtually loses its statistical association with Confidence in major companies when instrumentalized. This implies a rejection of Hypothesis 1. Yet from the test of exogeneityFootnote 15 of Social trust, given at the bottom of Table 2, we see that the latter can actually be treated as exogenous (p-value of 0.61), which leads us back to results contained in Table 1.
Table 2. Confidence in major companies: Bivariate Probit Estimates and Exogeneity Test for Social trust
Notes: *p < 0.10, **p < 0.05, ***p < 0.01; bivariate regression coefficients are shown; cluster-robust standard errors in parenthesis; clustering on country level with 55 clusters; control variables not shown for brevity; higher values of the regulation variables imply less strict regulations (Extent, Credit Market Regulation, Labor Market Regulation) and higher impartiality/quality of regulation (Impartiality).
As mentioned above, the original variable measuring confidence in major companies is recorded on a 4-point scale. To investigate the variation in this variable we apply ordered and generalized ordered probit estimators. The latter estimator is based on a series of individual probit regressions where the binary variable to be explained varies across estimations. Different combinations of the four possible answers are aggregated into value 1 of the binary variable. One advantage of this estimator over ordered probit is that it does not rely on the “parallel lines” assumption (see Williams, Reference Williams2016, for details).
Results are provided by Table 3. Column (1) gives the findings from the ordered probit estimator and the remaining columns form the generalized ordered probit estimates. The LR-test shown at the bottom of Table 3 signals that the parallel lines assumption underlying the ordered probit estimator is not fulfilled and that the generalized ordered probit results are preferred.
Table 3. Confidence in major companies: Generalized Ordered Probit Estimates
Notes: *p < 0.10, **p < 0.05, ***p < 0.01; generalized ordered probit regressions gives a series of probit regressions where the binary variable to be explained varies across the estimations (different combinations of answer possibilities); endogenous variable with entry 1 = high confidence; 2 = quite a lot confidence; 3 = not very much confidence; 4 = no confidence at all; regression (probit and ordered probit) coefficients are shown; cluster-robust standard errors in parenthesis; clustering on country level with 55 clusters; control variables not shown for brevity; higher values of the regulation variables imply less strict regulations (Extent, Credit Market Regulation, Labor Market Regulation) and higher impartiality / quality of regulation (Impartiality).
Social trust, Impartiality of business regulation and GDP growth are statistically significant with expected signs in each individual probit regression. Extent of business regulation and Unemployment rate are significant in two out of three regressions. When reply option “high confidence” is separated from the other three reply possibilities these two variables lose their statistical significance but keep the expected sign. Taken together, using the 4-point scale leaves substantive findings unaltered and confirms the validity of our hypotheses.Footnote 16
6. Summary and conclusions
Although firms are a key organization in capitalist market economies and public confidence in the business sector is crucial for well-functioning markets, there is surprisingly little empirical work on its sources. This study investigates statistical relationships of confidence in major companies with generalized social trust, macroeconomic performance, and quality and extent of business regulation. We base our investigation on a multilevel data set, employing individual-level data for generalized social trust and confidence in major companies and country-level data on economic regulation.
The central idea of the paper is that in addition to generalized social trust and good macroeconomic performance, business regulation can serve as a tool to improve confidence, as confidence in companies hinges on rules, formal procedures, and practices that shape how major firms behave. We consider two dimensions of business regulation: the qualitative (“process”) dimension is measured by the impartial and effective enforcement of regulatory provisions, and the quantitative dimension (“extent” or “intensity”) of regulation depends on the severity of regulatory constraints for market competition. We argue that it is not the intensity dimension of business regulation per se that matters for confidence in companies, but that the impartiality of procedures to enforce regulation is decisive.
Our findings are consistent with the notion that generalized social trust is a robust predictor of confidence in major companies. Its substantive impact, however, is limited, and supports the view that people perceive their relationship with major companies as hierarchical. Our empirical investigation also provides evidence that bad macroeconomic performance hampers confidence in major companies. Thus the idea that besides political institutions and actors, major market organizations are also blamed for a bad macroeconomic performance receives support from our analysis.
We show that the impartiality of business regulation matters to confidence in major firms. Our results imply that less strict business and credit market regulation are conducive to confidence in major companies. This finding is consistent with the view that market competition serves as a disciplining device and instills incentives to trustworthy behavior by market actors, which in turn impacts positively on confidence. It indicates that freedom of competition appears to be more important to the formation of confidence than more intensive regulatory provisions, as is already shown at a macro-level in the seminal paper of Berggren and Jordahl (Reference Berggren and Jordahl2006).
It is important to stress that the evidence provided should be seen as an indication of a very robust partial correlation between our variables of main interest rather than representing causal channels. While we try to cope with endogeneity concerns via instrumenting social trust, a battery of control variables including country and year fixed effects and using a mixed micro/macro-variable approach, we do not suggest that these measures are sufficient to allow for an interpretation of our findings in a causal sense. Omitted variables still cannot be ruled out, not least due to the impossibility of modeling individual specific fixed effects with the available EVS/WVS data set.
Nevertheless, taken together and bearing this caveat in mind, the results regarding the relationship of generalized social trust, regulatory policies, and confidence in major companies signal that institutional reformers do have some room for maneuver for improving public confidence in the business sector. Governments can help companies in their attempt to regain confidence to levels before the financial crisis of 2008 by the credible implementation of impartial business regulation, that is, regulation that is not perceived by the public as primarily serving narrow vested interests of specific groups.
Appendix
Table A1. Descriptive statistics
Table A2. Pairwise correlations (variables of main interest plus GDP per capita)
