Indian Labor Regulations and the Cost of Corruption Appendix

Appendix 1: Full Results by State and Industry46 Table 10: Estimates of Tau by State State 46 Tau Standard Error Andhra Pradesh -.159 Assam .322 Bihar .693 Delhi .427 Gujarat .165 Himachal Pradesh -.165 Haryana .007 Jharkhand .388 Karnataka .52 Kerala .138 Maharashtra .332 Madhya Pradesh .379 Orissa .283 Punjab .096 Rajasthan .32 Tamil Nadu .397 Uttar Pradesh .502 West Bengal .151 .038 .041 .069 .048 .047 .023 .044 .061 .06 .033 .038 .047 .044 .041 .05 .059 .069 .054 Standard errors generated using a wild cluster bootsrap procedure with 200 replications. Table 11: Estimates of Tau by Industry Industry Tau Standard Error Mining and quarrying Manufacturing Electricity, gas and water supply Construction Wholesale and retail trade Hotels and restaurants Transport, storage and communications Financial intermediation Real estate, renting and business activities Public administration and defence Education Health and social work Other service activities Extraterritorial organizations and bodies -.042 .268 -.367 .478 .637 .468 .334 -.105 .601 -.311 -.173 .076 .264 .024 .047 .039 .022 .047 .115 .06 .056 .044 .062 .031 .042 .03 .057 1.315 Table 12: Estimates of Tau by Owernship Type Ownership Type Tau Standard Error Government and public sector undertaking -.092 Non profit institution -.04 Unincorporated proprietary .43 Unincorporated partnership -.058 Corporate non financial -.197 Corporate financial -.18 Co-operative -.007 .028 .038 .087 .028 .026 .023 .022 Appendix 2: Anecdotal Evidence Regarding Harassment Bribery from “ipaidabribe.com” “I am a small factory owner in Kirti Nagar Industrial Area. We follow almost all rules laid down by government for the welfare of workers. Now, even if we follow everything there is always somethings where we lack and which needs improvement. We have a factory inspector by the name of Mr.R.B.Singh (M: 9818829355). He comes to all the factories in our area, inspects them, find mistakes and then harass and blackmails us. According to him he can get our factories sealed. To avoid this, to save our time and to save the unnecessary paperwork we pay him every year. I have paid him twice in two years i.e. 10000 & 15000 and this is common with all factories. Please take a strict action against him so that he learns a lesson. I am sure he is not alone. All his colleagues are equally corrupt.” (Reported on August 11, 2014 from New Delhi, Delhi — Report #131791) “During the routine labor verification process by the labor department at our office, we were advised by the consultant to pay the labor inspector a bribe to ensure that they don’t keep calling us for needless paperwork.” (Reported on June 28, 2011 from Chennai, Tamil Nadu — Report #35064) “The Labour Department requires a dozen odd registers to be maintained some of them which are totally outdated and pointless. E.g: Salary register, Attendance register, Leave register etc. Our IT office has an electronic system that logs all entries/exits and leave taken. We have the records and offerred to provide it to them in a printout. Salaries are paid electronically via bank transfer. The officer declined and said it must be maintained in a manual register! Finally an arrangement was made where we maintain a few records manually and the rest he would overlook. Cost of arrangement Rs 1500 twice a year even if the officer shows up only once a year for the inspection! He is supposed to inspect twice so expects to be paid even for the time he did not show up!” (Reported on October 13, 2010 from Chennai, Tamil Nadu — Report #44950) “Well i had gone to renew my labour license and after all the running around in the bank and the department, the signing authority asked me to pay Rs.500 for signing. When asked why 500, i was told since there are 5 employess for Rs.100 each.” (Reported on December 31, 2010 from Hyderabad, Andhra Pradesh — Report #43509 ) “... in my third visit i met one of office peon in Labour office he guided me for the bribe he also investigated and advised me for bribe according to the number of Employees deployed on contract basis and for this valueble suggestion he charged me Rs. 100. Again with full confidence i went to the ALCs desk and straight away i offered him the packet which was contains the amount of Bribe Rs. 3000/- ... He issued me the license after office hours...” (Reported on March 30, 2011 from Mumbai, Maharashtra — Report #39133) “Applying for shop & establishment [registration] & procured all documents relating to the registration. Finally inspectors are asking Rs.1000 as a bribe. If any other notice received by the company for resolving that another Rs.2000 and above , it depends on the company” (Reported on March 28, 2014 from Bangalore, Karnataka — Report #99016) “Officer name Naveen Kumar . Mobile no. 9468104694- He is asking for a bribe of 60,000 and is saying will issue a negative report under labour laws.” (Reported on January 24, 2014 from Gurgaon, Haryana — Report #83365) Appendix 3: Modeling Extortion (ie: Harassment Bribery) In this Appendix we model size-based regulations in an environment where corrupt inspectors use the fact that de jure regulations are numerous, complex and burdensome in order to extort bribes from firms (ie: harassment bribery). The model aims to illustrate how corruption may lead to higher per worker costs for firms that exceed the 10 worker threshold, while being as parsimonious as possible. The set-up and timeline is described below and in Figure 8. First, firms must choose their number of workers (n ≥ 10 or n0 < 10).47 As in Section 4.1, firms are characterized by a productivity parameter α, so that firms with higher productivity would like to choose higher n. If firms choose n greater than or equal to 10, they come under the legal purview of size-based regulations, which makes it more difficult for them to appeal extortionary practices on the part of inspectors. After choosing a level of employment, firms are randomly matched with an inspector. With probability κ, the inspector is corrupt; otherwise the inspector is honest. An honest inspector will enforce a reasonable interpretation of the spirit of the regulations if the firm has more than 10 workers. To be compliant with this “reasonable” interpretation of the regulations will cost the firm an amount F1 (n), which may in general depend on the number of workers in the firm. A firm with fewer than 10 workers incurs no regulatory costs if matched with an honest inspector. If the firm is instead matched with a corrupt inspector, the inspector will threaten to report the firm for technical infractions unless it pays a bribe (which we denote b or b0 , depending on whether the firm has chosen n ≥ 10 or n0 < 10), the value of which is 47 Throughout, primes will denote the values of variables on the side of the decision tree in which firms hire less than 10 workers. determined by Nash Bargaining. The firm may choose to pay the bribe or appeal the threatened fine in court. If appealing the fine in court, the firm will win with probability p (or p0 ) but will incur legal fees (cL ) with certainty. If it wins the case, the firm has no further financial obligations. If the firm loses, it is obliged to pay an amount F2 (n), which we take to be much larger than F1 (n). This last assumption is tantamount to supposing that a reasonable level of compliance with regulations is not extremely costly in comparison to the punishments that could be brought by an inspector for violating the regulations which may be a reasonable assumption in contexts where inspectors have a great amount of bargaining power and/or punishments can involve prison sentences. The assumption is also necessary for the framework to be one of extortion rather than collusion: if F1 (n) were large in comparison with F2 (n), firms would benefit from collusion and would face lower costs with corrupt inspectors than with honest ones. It is also plausible that both F2 (n) and F1 (n) are increasing functions of the number of workers, especially if we acknowledge that the full cost of any fine would include the opportunity cost of a manager’s time. We will consider the case where the total fines are directly proportional to the number of workers: Fi (n) = fi ∗ n. The decision tree representing the firm’s choices described above is provided in Figure 8. Figure 8: Decision Tree Firm n’ ¡ 10 n ≥ 10 Nature Nature honest inspector corrupt (prob κ) honest inspector corrupt (prob κ) π(α, n0 ) 0 Firm π(α, n) − F1 (n) 0 Firm pay appeal pay appeal π(α, n0 ) − b0 b0 Nature π(α, n) − b b Nature win with prob p0 lose win with prob p lose π(α, n0 ) − cL 0 π(α, n0 ) − cL − F2 (n0 ) 0 π(α, n) − cL 0 π(α, n) − cL − F2 (n) 0 An important assumption is that p0 , the probability of a firm’s winning the case when n0 < 10, is much higher than p, the probability of winning the case when n ≥ 10. The idea is that a firm with less than 10 workers is not under the legal purview of the regulations, so any case regarding regulatory infractions brought against the firm would have no standing in court. In what follows, we will take p = 0 and p0 = 1 for simplicity. As previously mentioned, if the firm meets a corrupt inspector, the value of the bribe paid to avoid going to court is determined through a process of Nash Bargaining over the surplus, where α and β are the bargaining weights of the inspector and firm, respectively: max (b)α (cL + (1 − p)F2 (n) − b)β b 0 )F2 (n)) 2 (n)) (and b0 = α(cL +(1−p , The solution of this maximization problem is that b = α(cL +(1−p)F α+β α+β for firms with less than 10 workers). Given that firms meet corrupt inspectors (and thus pay bribes) with probability κ and meet honest inspectors (and thus pay F1 (n)) with probability 1 − κ, the expected cost for a firm with greater than 10 workers is κb + (1 − κ)F1 (n), while the expected cost for a firm with less than 10 workers is κb0 . Taking the difference and substituting in our expressions for b and b0 , we get that firms that cross the 10 worker threshold α face an increase in expected costs of κ α+β (p0 − p)F2 (n) + (1 − κ)F1 (n). We are interested, however, in the increase in per worker costs that firms face when exceeding the 10 worker threshold, not the increase in total costs (as discussed earlier, an increase in per worker costs is the only way to produce a downshift in the logged firm size distribution in a static model). Thus, we divide the last result by the number of workers, n, to get per worker costs. Before doing so, we make the further simplifications that p = 0, p0 = 1, α and β both equal 1 (equal bargaining weights), and that all fines are proportional to firm size (Fi (n) = fi ∗ n). Then, the increase in per worker costs for firms that exceed the 10 worker threshold is κ f22 + (1 − κ)f1 . From the last result we see that if f2  f1 (in particular, in this case, if f2 > 2f1 ), then the increase in a firm’s per worker costs for exceeding the 10 worker threshold (i.e.: what we call τ in the paper) is increasing in the proportion of corrupt inspectors, κ. Again, that we are considering a context of extortion or “harassment bribery” is implied by the assumption f2  f1 . It is this condition (that f2 is very large) that gives corrupt inspectors the power to extract heavy bribes. We think it is a reasonable assumption given the anecdotal evidence regarding bribery we have found (some of which we present in Appendix 2). To conclude, the model above illustrates conditions that may explain the correlations we observed between corruption and τ in Section 6. In particular, the conclusion of the model is that firms in states with a higher proportion of corrupt inspectors (ie: more corrupt states), face higher per worker costs for exceeding the 10 worker threshold (higher τ ).