Enhancing or attenuating? How patronage ties moderate the impact of performance feedback on performance improvement in the public sector

ABSTRACT

This research investigates how performance feedback affects public organizations’ performance improvement and, more importantly, how patronage ties moderate the impact. Using China’s official city air quality ranking as the case, our empirical analyses with standard panel data models and the regression discontinuity design find that negative performance feedback leads to subsequent performance improvement and, more crucially and interestingly, that patronage ties enhance, rather than attenuate, this effect. Essentially, this study deepens and extends the performance feedback theory in the public sector by bringing in informal political institutions as boundary conditions, offering more nuanced understandings of organizational responses to performance feedback.

KEYWORDS:

• Performance feedback
• patronage ties
• organizational performance
• public organizations

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/14719037.2024.2389206.

Notes

1. Of note, this particular strategy of ‘avoiding blame by doing good’ can be understood as a type of policy strategy according to Hood’s categorization of BA strategies (Hood 2010). The rationale underlying this strategy is as follows: on the one hand, if public organizations manage to improve their performance to a level that surpasses the aspirations, they address the blame for the performance failure fundamentally; one the other hand, performance improvement itself signalizes responsiveness to organizational problems, which helps improve organizations’ reputation and thus mitigate external blame for performance shortfalls (Holm 2018).

2. It is worth noting that patronage ties in democratic and non-democratic regimes may take different forms in terms of what patrons and clients exchange (Paik and Baum 2014). For instance, a typical form of patronage relations inhibiting in party politics in democratic regimes is that politicians as patrons distribute public resources (e.g. public employment) in exchange for electoral support from the clients. However, in non-democratic regimes without competitive election, this form of patronage ties does not exist.

3. The negative impact of clients’ poor performance on the higher-up authorities’ evaluation towards the patrons could be especially evident when patrons rely on their clients to perform well in the first place to put pressures on other agents to follow suit as previously noted.

4. See the appendices (Part F) for an illustrative example.

5. The coverage of ranked cities was 169 at first. However, a city (Laiwu) was incorporated into another city and removed from the ranking since January 2019, making the final number 168.

6. It is also common for patronage ties to form through political appointments in Western democracies, in which politicians (as patrons) distribute public employment in exchange for support from the bureaucratic clients (Toral 2022).

7. Detailed discussions on the potential endogeneity caused by cities’ prioritization of environmental protection are provided in the appendices (Part C).

8. Analyses using other approaches, such as the Newey-West approach and FGLS (feasible generalized least squares) estimations, were also performed to obtain heteroskedasticity- and autocorrelation-consistent standard errors (Wooldridge 2010). Results are largely consistent with those using clustered robust standard errors.

9. In the sharp RD design, the probability of treatment from 0 to 1 at the cut-off; in the fuzzy RD case, the probability jumps by less than one.

10. See more detailed discussions on this issue in the appendices (Part A).

11. Econometric specifications are provided in the appendices (Part B).

12. Due to space limitations, discussions on results for control variables are presented in the appendices (Part D).

13. Further discussions on the potential reasons for why the enhancing effect dominates in our empirical context can be found in the appendices (Part E).

14. We provide preliminary discussions on this issue in the appendices (Part E).

Additional information

Funding

This work was supported by the [Social Science Fund of Hunan Province of China] under Grant [19ZDB006].