One distinct feature of the rise in income inequality over recent decades is the surging incomes of the working rich, particularly the pay of a class of top managers. A popular view stresses the role of performance-related pay in generating this phenomenon. According to this view, the pay of many top managers is high-powered – with incentives tied to firm performance. In good times, managerial pay is boosted by firm profits, while in bad times top managers are barely punished because of limited liability. Such an incentive structure generates excess pay to top managers, dwarfing the pay to salaried workers.
While being accused for unduly rewarding the managerial class, incentive pay has long been recognised as an important sorting mechanism to allocate talents to jobs. For example, Lazear’s (1986) well-known theory posits that high-ability workers sort into jobs with performance pay while low-ability workers take jobs with fixed wages. Along these lines, inequality caused by incentive pay is justifiable to the extent that labour sorting affects economic efficiency.
In a recent paper (Wu, 2017), I develop a theory to assess the role of incentive structure in sorting labour to achieve efficiency on the one hand, and in generating wage inequality on the other. The theory links worker ability, pay-performance sensitivity, and pay levels across a wide range of jobs. In short, I integrate a principal-agent problem into a Lucas (1978) general equilibrium framework. The theory starts from a distribution of talent in an economy in which individuals differ in their ability to manage firms. To materialise their ability, an individual needs to be hired by a firm. The firm designs incentive contracts to elicit managerial effort which enhances firm productivity. But the design of incentive contracts is subject to limited liability in the sense that the firm cannot punish its manager beyond a certain limit, even when the manager fails to improve the firm’s profitability. Market competition determines the allocation of individuals to firms mediated by employment contracts, the distribution of firm size, and ultimately the distribution of wage income in the economy.
Sorting of managers into incentive contracts
The central result derived from the theory is the sorting of individuals, on the basis of managerial ability, into three types of employment status – production workers, small business owners, and salaried managers – corresponding to three basic types of incentive contracts: fixed salary, residual claim, and contingent pay. The class of small business owners emerge because of a well-known result in contract theory – when the surplus generated by the employment relationship is sufficiently small, the firm owner can sell the ‘store’ upfront to its manager to extract all the surplus. This ownership transfer is an efficient contractual arrangement with little cost. When the surplus is large and ownership transfer is not feasible, providing incentives to managers involves a structure of contingent pay and requires a cost of paying above managers’ outside options. However, for a medium-talent manager, the value of managerial effort is not sufficient to outweigh the cost of providing a high level of incentive. Thus, the owner optimally uses a contingent-pay structure tied to the manager's outside option. This type of contract is relatively low cost and has modest benefits. By contrast, a high-talent manager creates a sufficient surplus such that the owner is willing to offer a rent-sharing contract – giving part of the surplus to managers. Such a rent-sharing contract incurs a high cost but has a large benefit.
The sorting pattern described above produces a relationship between incentive structure and pay level that is not only consistent with prior studies of top managers, but also sheds light on the wage distribution among lower-level managers and other workers. The first implication concerns the divide between managers and workers. Managers as a whole earn a wage premium over workers as the result of (1) the selection effect that allows managers to utilise their managerial ability, (2) incentive contracts used to elicit managerial effort, and (3) firm size, which amplifies returns to managerial ability. Using the US Occupational Compensation Survey data, I find evidence consistent with these three drivers of the managerial premium.
The second implication concerns the wage distribution across individuals with heterogeneous managerial ability. In the theory, a range of less-talented individuals with different abilities are rewarded by different incentive structures but receive a similar level of pay. An alignment between incentive structure and pay level occurs only for high-talent managers, who manage large businesses and also share their employers’ profits. These predictions are borne out by the US National Compensation Survey (NCS) data. Figure 1 below plots the hourly wages against the ‘managerial level’ defined in the NCS data. Notably, there is a convex relationship between wages and managerial levels. Moreover, as a measure of pay dispersion, the difference between the 10th and 90th percentile wages at the same managerial level widens as the managerial level rises.
Figure 1 Distribution of hourly wages across managerial levels
Notes: The hourly wages are the mean wage at each managerial level across all occupations from 2006 to 2010 in the US National Compensation Surveys. The managerial level is defined by the skill and responsibility for an occupation, ranked from Level 1 to Level 15. Levels 1 to 5 are mostly related to non-managerial work.
Data source: US Bureau of Labor Statistics.
Technology progress and wage inequality
The theory I develop offers a new perspective for studying the effects of technological progress on both the wages and employment of various groups of individuals. One leading explanation for the rising inequality between unskilled and skilled workers is skill-biased technology progress, which disproportionately increases the demand for skilled workers (e.g. Piketty and Saez 2003, 2006, Acemoglu and Autor 2012). I show that skill-neutral (i.e. TFP-enhancing) technological progress can generate a substantial impact on wage inequality through the distribution of firm size. In particular, TFP-enhancing technological progress causes disproportionate growth in firm size favouring more-productive firms and also increases the value of managerial efforts in big firms. This induces the reallocation of resources from small firms (run by less talented managers) to big firms (run by more talented managers). In consequence, three effects occur: (1) the shrinkage of small business owners, (2) an expansion of high-talent managers who share firm profits, and (3) an increase in the level of incentives offered to managers. These three effects jointly contribute to a highly skewed wage distribution in favour of top talent.
These predictions are consistent with broad data patterns regarding the wage dynamics and long-term employment in the US. The top panel of Figure 2 plots the managerial wage premium – defined as the difference in the annual real (CPI-adjusted) wages between managers and production/clerical workers – over the period 1997-2014. The wage premium for salaried managers rises from approximately $20,000 in the 1990s to more than $30,000 in the 2010s. The premium for top managers (CEOs and general managers) more than doubles during the sample period, whereas the premium for mid-level managers (division managers and plant managers) rises far more modestly. Isolating low-level managers (first-line supervisors and administrative managers) shows that their wage premium barely changes over time.
Figure 2 Managerial premium and employment over time
Notes: The difference in real wage is defined as the difference between the annual average wage of a particular type of managers (defined by the description of their occupations) and that of production and clerical workers, normalized by yearly national CPI. Low-level managers are ‘first-line supervisors’ or ‘administrative managers’. Higher-level managers are ‘general managers’ or chief executives. Mid-level managers are individuals whose occupations belong to the managerial category, excluding the low-level managers or higher-level managers. Self-employment are individuals classified as “Self-employed” in the US Labor Statistics.
Data source: US. National Occupational Compensation Statistics, US Bureau of Labor Statistics.
The bottom panel of Figure 2 shows that the proportion of the self-employed (a proxy for small business owners) in total employment gradually decreases over the 2003-2011 period, and that the proportion of salaried managers exhibits a similar downward trend. Moreover, the share of mid-level managers among all managers (self-employed and salaried managers combined) increases significantly over time, although the increasing trend in the share of higher-level managers is less obvious.
The theory posits a trade-off between wage inequality and economic efficiency. Changes in market conditions and regulation that improve economic efficiency inevitably increase wage inequality. I discuss some policy implications derived from this insight.
- Product market competition. The theory predicts that product competition contributes to a highly skewed wage distribution in favour of top talent. This is because stiffer competition triggers production factors to be reallocated from smaller to larger firms. As a result, larger firms offer greater incentives to their managers, while smaller firms and less talented managers exit the market. Thus, both efficiency and inequality increase. This prediction has implications about policies regarding international trade. In particular, an economic sector that is subject to more intense import competition should witness greater income inequalities among workers.
- Limiting mobility of managers. An extension of the theory shows that when the mobility of managers increases, firms will compete more intensively for top managers, whose bargaining power to share firm profits then increases. In consequence, the distribution of managerial pay becomes more skewed. This suggests that one policy that may curb the growing wage inequality is to limit the mobility of managers across firms. But this policy may hurt economic efficiency to the extent that managerial slack is exacerbated.
- Enhancing managerial liability. Is it possible to obtain more efficiency and less inequality at the same time? According to the theory, one such possibility is to enhance managerial liability, that is, to allow firms to punish managers when they fail to improve firm profitability. In this way, the firms can align managerial incentives with firm objectives without giving away too much rent to managers. In an extreme case, the firms can use a contract analogous to ‘selling the store’ to managers but extract all the surplus upfront. Then, economic efficiency is improved, firms keep most surplus, and managers’ wages are largely disciplined by their outside options.
Acemoglu, D and D Autor (2012), “Skills, Tasks, and Technologies: Implications for Employment and Earnings”, in O Ashenfelter and D Card (eds), Handbook of labour Economics, Vol. 4b, Elsevier Science North Holland.
Lazear, E (1986), “Salaries and Piece Rates”, Journal of Business 59(3): 405-431.
Lucas, R Jr. (1978), “On the Size Distribution of Business Firms”, The Bell Journal of Economics IX: 308-523.
Piketty, T and E Saez (2003), “Income Inequality in the United States, 1913-1998', Quarterly Journal of Economics 118(1): 1-39.
Piketty, T and E Saez (2006), “The Evolution of Top Incomes: A Historical and International Perspective”, American Economic Review 96(2): 200-205.
Wu, Y (2017), “Incentive Contracts and the Allocation of Talent”, Economic Journal 127(607): 2744-2783.