Do Inequality-based Entry Barriers Deter The Formation of Female-Owned Firms In Nigeria?

Transcription

1 Current Verson: December 30, 2013 Do Inequalty-based Entry Barrers Deter The Formaton of Female-Owned Frms In Ngera? Tffany R. Bussey * Julet U. Elu ** Gregory N. Prce *** Abstract In ths paper, we consder the role of nequalty-based entry barrers on the formaton of female-owned frms n Ngera. Wth data from the 2010 World Enterprse Survey, we estmate the parameters of a smple model of female-owned frm entry to determne the role of nequalty-based barrers on the number of female-owned frms across cty-ndustry clusters n Ngera. Parameter estmates from count data specfcatons of frm entry reveal that access to fnancng, land, and lcenses/permts absolutely deter the entry of female-owned frms, as these entry barrers are proportonal to the probablty of observng no female-owned frms. In general, barrers to securng land constran the entry of female-owned frms beyond the process determnng absolute entry deterrence. Ths suggests that the market entry and underrepresentaton of female-owned among frmowners and entrepreneurs n Ngera s, at least n part, caused by gender nequalty n general. As prvate frm output domnates the gross domestc product of modern economes, our fndngs suggest that the reducton of gender nequalty n Sub-Saharan Afrca would result n more female-owned and entrepreneurs whch would catalyze economc growth. JEL Classfcaton: C25, J16, L1, L26, O55, Z13 * Morehouse Entrepreneurshp Center, 830 Westvew Dr. SW, Atlanta GA, emal: trbussey@morehouse.edu, Tel #: (404) * * Department of Economcs, Morehouse College, 830 Westvew Dr. SW, Atlanta GA, emal: jelu@morehouse.edu, Tel #: (404) *** Department of Economcs, Morehouse College, 830 Westvew Dr. SW, Atlanta GA, emal: gprce@morehouse.edu, Tel #: (404) Ths paper was prepared for presentaton at the 2014 Alled Socal Scences meetngs n Phladelpha, PA January 3-5 1

2 I. Introducton Gender nequalty n Sub-Saharan Afrcan economes has been found to be harmful to human development (Elu, 2013), and one of ts determnants economc growth (Balamoune-Lutz and McGllvray, 2009). 1 The recent fndng by Elu (2013) that the elastcty of human development n Sub-Saharan Afrca s a functon of female access to bank loans suggests that gender nequalty n Sub-Saharan Afrca could also constran the formaton of female-owned frms, as credt fnancng s a crucal nput for frms as ther s evdence that relatve to maleowned frms, female-owned frms n Sub-Saharan Afrca are credt-constraned (Asedu et.al, 2013). As frms and entrepreneurs enterng a market contrbute to an economy s gross domestc product, partcularly n emergng/transton economes (McMllan and Woodruff, 2002), and can even possbly cause economc growth (Audretsch, 2007; Koellnger and Thurk, 2012), gender nequalty that engender barrers to the formaton of female-owned frms could constran both economc growth and human development. In ths paper, we consder the role of nequalty-based entry barrers on the formaton of female-owned frms n Ngera. Wth data from the 2010 World Enterprse Survey, we estmate the parameters of a smple model of female-owned frm entry to determne the role of nequalty-based barrers on the number of female-owned frms across cty-ndustry clusters n Ngera. Surely, gender nequalty n the formaton of frms and entrepreneurshp s a concern n all of Sub-Saharan Afrca (Hallward-Dremeer, 2011). However, our focus on Ngera s motvated by the fact that as Afrca s largest country t has the most females approxmately 80.2 mllon grls and women and ts human development ndcators are worse than those of comparable lower mddle-ncome countres (Brtsh Councl of Ngera, 2012). As such, Ngera wth ts hgh female populaton and low human development provdes perhaps an deal settng to dentfy the effects of gender nequalty n Sub-Saharan Afrca on female-owned frm formaton, whch s a potental drver of economc growth. II. A Smple Lmt Proft Model of Frm Entry If frm owners are proft maxmzers, a smple parsmonous theory of entry can proceed from vewng a typcal frm as basng ts entry decson on whether or not t can earn a proft sustanable aganst noncompettve ncumbents wth cost advantages. Incumbents are assumed to possess advantages, assocated wth for example, superor producton technology, manageral skll, networks, that enables them to produce a level of output 1 Internatonally, Human Development s offcally measured by the Unted Natons Development Program s Human Development Index (HDI). The HDI s a composte ndex based on three sub-ndexes, one of whch s an ndex for a country s Gross Domestc Product (Anand and Sen, 2000) whch provdes a lnkage between human development and economc growth. 2

3 where prce s close to margnal cost to the dsadvantage of new entrants. Ths s the classc lmt proft model of Orr (1974) where the lmt rate of proft s defned as ncumbent proft consstent wth the level of entry barrers n the market/ndustry. 2 As the lmt rate of proft s unobservable, t wll be a functon of factors that mpose cost dsadvantages entry barrers on new entrants such that f they entered, they could not compete wth ncumbents on prcng and output. In a lmt proft model, frm entry n market/ndustry s determned by E = θ ( π - entry n market/ndustry, π s expected post-entry proft, * π ), where E s frm * π s the lmt rate of proft, and θ s the entry response parameter. As the lmt rate of proft s unobservable, followng Ban (1956), assume that there are J > 0 exstng entry barrers that condton the ablty of ncumbents to engage n prcng/output that deters new * entrants lnearly as π = α o + J α j= 1 j effects on the lmt rate of proft gven by and α j, and B j, where B j s entry barrer j n market/ndustry, wth margnal α o s a constant. Gven * π = α o + J α j= 1 j B j, t follows that n a market where ncumbents are at least monopolstcally compettve, but not monopoly, E / π = θ > 0, E / and decreases wth respect to ncreases n entry barrers. B j < 0, or new frm entry ncreases wth respect to ncreases n expected proft, The mplcatons of the smple lmt proft model of entry suggest an emprcal strategy for dentfyng the effects of nequalty-based entry barrers on the entry of female-owned frms n Ngeran goods/servce markets that are at least monopolstcally compettve but not monopoly. Gven measurements on the entry of femaleowned frms and nequalty-based entry barrers that mpose costs on new entrants relatve to ncumbents, the effects of these type of entry barrers on the market entry of female-owned frms n Ngera can be dentfed emprcally. III. Data To estmate the parameters of an emprcal Lmt proft model of entry for female-owned frms, we use data from the 2010 World Enterprse Survey for Ngera (WES). 3 The WES, a frm-level survey, has been admnstered by the World Bank snce Admnstered to busness owners and top managers of formal regstered frms wth 5 or more employees n face-to-face ntervews, the WES captures data from respondents on frm characterstcs, gender partcpaton, access to fnance, annual sales, costs of nputs/labor, workforce 2 For the emprcal evdence n support of the mplcatons of a lmt proft model of entry see Arauzo-Carod and Segarra-Blasco (2005), Fenberg (2010), Kheman and Shapro (1986), and Prce (1995). 3 WES data are avalable at: 3

4 composton, brbery, lcensng, nfrastructure, trade, crme, competton, capacty utlzaton, land and permts, taxaton, nformalty, busness-government relatons, nnovaton and technology, and performance measures. Over 90 percent of the questons objectvely measure characterstcs of a country s busness envronment. The remanng survey questons capture the respondent s opnon on the obstacles to frm growth and performance whch we explot to construct nequalty based frm entry barrers that can potentally deter the entry of female-owned frms. The samplng methodology for the WES s stratfed random samplng, and the strata are frm sze, ndustry/sector, and geographc regon wthn a country. Frm sze levels are 5-19 (small), (medum), and frms wth 100+ employees (large-szed frms) whch are oversampled. Industry/sector categores are typcally manufacturng, retal, and other servces, and for larger economes addtonal sub-ndustres/sectors are selected as addtonal strata on the bass of employment, value-added, and total number of establshments fgures. Gven the dfferent probabltes of selecton across the strata, WES provdes samplng weghts whch enable consstent estmaton of populaton parameters of nterest. Gven that the core varable of nterest s the market entry of female-owned frms n Ngera, we use the reported number of female frm-owners n a partcular cty-ndustry cluster as our dependent varable. Whle ths count varable s not necessarly a measure of new frm entry, t s an approxmaton to new frm entry, as the number of frms n a cty-ndustry cluster wll be proportonal to the number of new entrants at a pont n tme and the data are cross-sectonal across cty-ndustry parngs. We measure female-owned frm counts at the cty-ndustry cluster level to control for any unobservables assocated wth the tendency of frms to be attracted to the agglomeraton and localzaton benefts of partcular ctes and ndustres (Barbosa, Gumaraes and Woodward, 2004; Belleflamme, Pcard and Thsse, 2000; Glaeser et.al, 1992). Gven that the WES s a cross-secton, renderng t mpossble to use a measure of lagged ncumbent characterstcs, we use average proft of frms n each cty-ndustry cluster as a measure of expected proft. We construct four measures of entry barrers. The frst one s a measure of ncumbent frm sze, whch n a lmt proft model captures the ablty of an ncumbent frm to deter entry as a result of cost advantages. Our measure of ncumbent frm sze SCALE s the percentage of frms n a cty-ndustry cluster wth 100 or more employees. The remanng three entry barrers are based on the respondent s provded opnon on partcular obstacles to frm growth and performance. In partcular we measure nequalty-based entry barrers based on the percentage 4

5 of ncumbents n a cty-ndustry cluster who reported that the most serous obstacle to frm growth/performance s access to fnance (FINANCE BARRIER), access to land (LAND BARRIER), and securng busness lcenses/permts (REGULATORY BARRIER). To the extent that relatves to males, these barrers are hgher for females, they consttute nequalty-based entry barrers that deter entry as they are a source of relatve cost advantages for ncumbents. IV. Results Table 1 reports a covarate summary. On average across the 179 cty-ndustry clusters, the average number of female-owned frms s less than one, underscorng a severe underrepresentaton of female-owned formal frms n Ngera. Among the measured entry barrers, the hghest average value s for access to Fnance, suggestng that securng fnancng s relatvely dffcult for all formal frms n Ngera. Regulatory barrers appear to be low relatve to other entry barrers, suggestng that the regulatory burden for formal frms, on average s relatvely low n Ngera. Our strategy for estmatng the parameters of an emprcal Lmt proft model of female-owned frm entry recognzes that the number of female-owned frms n a cty-ndustry cluster s dscrete, nteger-valued, and consttutes count data. As such, we assume that the number of female-owned frms are realzatons from ether a Posson, or, f there s unobserved heterogenety, a Negatve Bnomal count dstrbuton. We specfy and estmate the parameters of relevant Posson and Negatve Bnomal regresson specfcatons of a Lmt proft model of female-owned frm entry, where the estmated parameters measure the effects of exogenous entry barrer varables on λ the expected value of the number of female-owned frms n a cty-ndustry cluster. 4 4 A Posson regresson model (Cameron and Trved, 1998) s formulated by specfyng for nteger-valued measures of female-owned frms E n cty-ndustry cluster, the condtonal mean λ as : ' lnλ = β Θ where β s a coeffcent vector, and Θ s a vector of exogenous expected proft and entry barrer measures that determne the expected value of the number of female-owned frms E n the th cty-ndustry cluster. The loglkelhood functon L( β ) has a gradent and Hessan gven by: L β ) = β ( ' β ' Θ [ Θ ( E e )] = 0 5

6 Table 2 reports Posson and Negatve parameter estmates of the Lmt proft model of female-owned frm entry n Ngera. 5 As there are a preponderance of zero female-owned frms across the cty-ndustry cluster groupngs n the sample, the probablty of zero female-owned frms n a gven cty-ndustry cluster may be too hgh to be captured adequately by smple Posson and Negatve Bnomal specfcatons. 6 As such, we also report the parameters of Zero-nflated Posson (Lambert, 1992) and Negatve Bnomal (Mullahy, 1986) specfcatons, whch explctly account for the process determnng the zero counts. The Zero-nflated specfcatons are potentally nformatve n ther own rght. As they account for the process determnng zero counts, they can nform how the market entry of female-owned frm s absolutely deterred by gender nequalty, as captured n the nequalty-based entry barrers. As for the adequacy of the specfcatons reported, Table 2 also reports the value of the Akake Informaton Crteron (Akake, 1998). Explct hypothess testng on the mpled restrctons of each specfcaton estmated, do not nform on how well the estmated model approxmates the true model whch s arguably the ultmate goal of statstcal nference. The Akake Informaton Crteron (AIC) s a measure of the nformaton dscrepancy between the 2 L( β ) = ' β β ' [ ( E E ) e β ' Θ )] < 0 Equatng the gradent to zero solves for β, and the negatvty of the Hessan ensures a global maxmum of the log-lkelhood estmator of the coeffcents n β. As a Posson specfcaton assumes there s no unobserved heterogenety, the mean and varance of λ are dentcal. Gven the possblty of unobserved heterogenety, the Posson model can be modfed as a Negatve Bnomal (Cameron and Trved, 1998) where the specfcaton of λ s: ln ' λ = β Θ + ε where ε reflects unobserved heterogenety causng the mean and varance of λ to dffer. 5 All estmates are weghted wth the cross-product of the sze, regon and establshment stratum weghts, and the standard errors are clustered on cty-ndustry groupngs to control for the unobservables assocated wth localzaton and agglomeraton. 6 In the sample, approxmately 74 percent of the cty-ndustry clusters have zero female-owned frms. Ths preponderance of zero counts could mmc overdsperson dfferent values for the mean and varance of λ nducng false selecton of a Negatve Bnomal specfcaton as a result of falng to account for the process determnng the zero counts. In a zero-nflated count specfcaton, the realzatons come from two regmes. In one regme the outcome s always zero, whle n the other t s not always zero. Let ω be the probablty that a realzaton has a zero outcome, then for a Posson dstrbuton, the mean s (1 - ω ) β ' Θ, where ω = [ β ' Θd ]/[ 1 + β ' Θd ], s a Logt specfcaton, and Θ and Θ d could be dentcal. A Zero-nflated Negatve Bnomal specfcaton smply adds a stochastc error term to the specfcaton for the condtonal mean. 6

7 estmated model and the true populaton model. The smaller the AIC for an estmated model, the lower s the dscrepancy between t and the true populaton model. Gven the AIC values across the estmated specfcatons of the Lmt proft model of female-owned frm entry n Table 2, the mnmum value s acheved for the Zero-Inflated Posson specfcaton, suggestng ts adequacy relatve to the other specfcatons. Consstent wth a Lmt proft model of entry, the statstcally sgnfcant parameter estmates suggest that n Ngera, the entry of female-owned frms n Ngera ncreases wth expected proftablty, and decreases wth respect to ncreases n entry barrers assocated wth securng land and lcenses/busness permts. Whle the sgn on the entry barrer assocated wth securng fnancng s statstcally sgnfcant, the sgn s postve, whch runs counter to the expectaton of the theory underlyng the Lmt proft model of entry. However, for the Zero nflaton parameters from the Logt determnng the zero count regme, the barrers assocated wth securng fnancng conforms, as ncreases n entry barrers assocated wth securng fnancng have a statstcally sgnfcant and postve effect on the probablty that there wll be zero female-owned frms enterng the market. In general, all the zero-nflaton parameters are statstcally sgnfcant and cohere wth the theory of the Lmt proft model. Ths suggests that n Ngera, gender nequalty n access to fnancng, land, and n the ablty to secure lcenses/permts results appears to absolutely deter the market entry of female-owned frms. The parameter estmates of the effects nequalty-related entry barrers from Zero-nflated Posson specfcaton are also of practcal sgnfcance. Wth respect to the regulatory barrers assocated wth securng land, the parameter estmate has an elastcty, evaluated at the means of all regressors of approxmately 8 percent. Ths mples that f on average across all 179 cty-ndustry clusters, barrers to securng land fell by 50 percent, approxmately 5 new female-owned frms would be nduced to enter the market, producng addtonal net output worth approxmately 36.7 mllon Nara whch s equal to approxmately $227,540 as of June 28, As there are a total of 47 female-owned frms n the sample, the entry of 5 consttutes an approxmately 11 percent ncrease n the number of female-owned frms. Smlar results follow from consderng a reducton n the entry barrers from the Zero-nflated parameter estmates, and underscore how practcally sgnfcant reductons n the entry barrers faced by female-owned frms are as the majorty of cty-ndustry clusters have no female-owned frms at all. V. Concluson 7 The value of the addtonal net output follows from multplyng average frm proft reported n Table 2 by the addtonal number of female-owned frms that would enter the market as a result of a 50 percent reducton n barrers to securng land. 7

8 Ths paper, consdered the role of nequalty-based entry barrers on the formaton of female-owned frms n Ngera. Parameter estmates from count data specfcatons of the number of female-owned frms across ctyclusters revealed that access to fnancng, land, and lcenses/permts absolutely deter the entry of female-owned frms, as these entry barrers are proportonal to the the probablty of observng no female-owned frms. In general, barrers to securng land constran the entry of female-owned frms beyond the process determnng absolute entry deterrence. Our results suggest that the market entry and underrepresentaton of females among frm-owners and entrepreneurs n Ngera s, at least n part, caused by gender nequalty n general, as the entry barrers capture the access of actual and/or potental female-own frms to fnance, land, and lcenses/permts requred to operate. As frm formaton s a potental drver of economc growth n Sub-Saharan Afrca, gender nequalty n the formaton of frms could constran economc growth, as Afrcan economes lose the output that could be produced by female-owned frms due to the entry deterrent effects on nequalty-based entry barrers. Our fndngs suggest that the reducton of gender nequalty n Sub-Saharan Afrca would result n more female frm-owners and entrepreneurs whch would catalyze economc growth. 8

9 Table 1 Covarate Summary Covarate Defnton a X σˆ N Number of Female-owned Number of frm owners n Frms n Cty-Industry cluster who ndcated they Cluster are female (queston = ng_b3f) Average Year 2008 Total Revenue - Total Cost: Cluster Proft Derved from the resdual In Nara of frm total revenue(queston = l1b) ( EXPECTED PROFIT) and the cost of land (queston = ng_f1a4b), cost of fnance (queston = ng_f1a122), cost of raw materals and ntermedate goods (queston = l2a), cost of labor (queston = l2b), deprecaton (queston = l2b), and rental cost of land, buldngs, equpment and furnture (queston = l2d) Percentage of Frms Bnary varable equal to n Cluster Wth f the frm has 100 or more or more employees 100 or more Employees(queston = sampsze) ( SCALE) Percentage of Cluster Bnary varable equal to Frm-Owners who report f a frm-owner ndcates that access to fnance s that access to fnance the most serous obstacle s the most serous obstacle ( FINANCE BARRIER) (queston = flb1) Bnary varable equal to Percentage of Cluster Frm-Owners who report f a frm-owner ndcates that access to land s that access to land the most serous obstacle s the most serous obstacle ( LAND BARRIER) (queston = flb1) Bnary varable equal to Percentage of Cluster Frm-Owners who report f a frm-owner ndcates that securng lcenses that securng lcenses and/or busness permts and/or busness permts s the most serous obstacle s the most serous obstacle(queston = flb1) ( REGULATORY BARRIER) Source: 2010 World Enterprse Surveys, ( Notes: a The queston ndcator refers to the partcular queston n the 2010 World Enterprse Surveys from whch the covarate s derved. X, σˆ, and N denote the mean, standard devaton, and number of cty-ndustry cluster observatons respectvely. 9

10 Table 2 Count Data Parameter Estmates: Entry Barrers and The Number of Female-owned Frms In Ngera Specfcaton: Posson Negatve Zero-Inflated Zero-Inflated Bnomal Posson Negatve Bnomal Regressand: Number of Femaleowned frms n cty-ndustry cluster Regressors: CONSTANT (.097) a (.112) a (.051) a (37.57) b EXPECTED PROFIT -5.15e e e e-09 (1.87e-09) (4.31e-09) (5.16e-09) b (2.07e-08) SCALE (3.95) b (5.50) c (10.0) c (10.28) b FINANCE BARRIER (2.86) c (3.65) (1.04) b (22.37) c LAND BARRIER (10.51) c c (10.0) (4.87) (11.36) REGULATORY BARRIER (1.74) a (2.40) a (.831) a (59.03) b Zero Inflaton parameters: Regressand: Bnary ndcator for Zero Female-owned frms In Cluster Regressors: CONSTANT (.444) a (.234) b EXPECTED PROFIT 5.25e e-09 (6.08e-09) a (1.70e-09) SCALE (9.89) a (1.96) FINANCE BARRIER (3.27) a (.914) b LAND BARRIER (12.40) a (1.52) REGULATORY BARRIER (3.60) c (3.11) Number of Observatons Akake Informaton Crteron Notes: Clustered standard errors n parentheses. a Sgnfcant at the.01 level b Sgnfcant at the.05 level c Sgnfcant at the.10 level 10

11 References Akake, Hrotugu "Informaton Theory and an Extenson of the Maxmum Lkelhood Prncple," n Selected Papers of Hrotugu Akake, pp , Sprnger, New York NY. Anand, Sudhr., and Amartya Sen The Income Component of the Human Development Index," Journal of Human Development 1: pp Arauzo-Carod, Josep-Mara, and Agustí Segarra-Blasco The Determnants of Entry are not Independent of Start-up Sze: Some Evdence From Spansh Manufacturng," Revew of Industral Organzaton, 27: pp Asedu, Elzabeth., Isaac Kalonda-Kanyama, Leonce Ndkumana, and Akwas Nt-Addae Access to Credt by Frms n Sub-Saharan Afrca: How Relevant Is Gender?," Amercan Economc Revew 103: pp Audretsch, Davd B Entrepreneurshp Captal and Economc Growth," Oxford Revew of Economc Polcy, 23: pp Ban Joe, S Barrers to New Competton." Harvard Unversty Press, Cambrdge MA. Balamoune-Lutz, Mna., and Mark McGllvray Does Gender Inequalty Reduce Growth n Sub- Saharan Afrcan and Arab Countres?" Afrcan Development Revew 21: pp Barbosa, Natala., Paulo Gumaraes, and Douglas Woodward Foregn Frm Entry n an Open Economy: The Case of Portugal," Appled Economcs, 36: pp Belleflamme, Paul., Perre Pcard, and Jacques-Franços Thsse An Economc Theory of Regonal Clusters," Journal of Urban Economcs, 48: pp Cameron, A. Coln and Pravn K. Trved Regresson Analyss of Count Data, Cambrdge Unversty Press, Cambrdge, UK. Elu, Julet U NEA Presdental Address: Gender Inequalty and Human Development n Sub-Saharan Afrca," Revew of Black Poltcal Economy, 40: pp Fenberg, Robert M Prce Impacts of Small-Frm Entry n US Manufacturng," Eastern Economc Journal. 36: pp Glaeser, Edward L., Hed D. Kallal, José A. Schenkman, and Andre Shlefer Growth n Ctes," Journal of Poltcal Economy, 100: pp Hallward-Dremeer, Mary Strengthenng Women s Entrepreneurshp," The Afrca Compettveness Report, pp , World Bank, Washngton DC. Kheman, R. S., and Danel M. Shapro The Determnants of New Plant Entry n Canada," Appled Economcs, 18: pp

12 Koellnger, Phlpp D., and A. Roy Thurk Entrepreneurshp and the Busness Cycle," Revew of Economcs and Statstcs, 94: pp Lambert, Dane Zero-nflated Posson Regresson Wth An Applcaton To Defects n Manufacturng," Technometrcs, 34: pp McMllan, John., and Chrstopher Woodruff The Central Role of Entrepreneurs n Transton Economes," Journal of Economc Perspectves, 16: pp Mullahy. John Specfcaton and Testng of Some Modfed Count Data Models," Journal of Econometrcs, 33: pp Orr, Dale The Determnants of Entry: A Study of the Canadan Manufacturng Industres," Revew of Economcs and Statstcs, 56: pp Prce, Gregory N The Determnants of Entry for Black-Owned Commercal Banks," Quarterly Revew of Economcs and Fnance, 35: pp Brtsh Councl of Ngera Gender n Ngera Report 2012: Improvng the Lves of Grls and Women n Ngera, Issues, Polces, Acton, 2nd edton, Abuja, NG. 12