ABSTRACT
In many countries the informal sector is a vital source of employment and income. But little is known about the impact of this sector on savings, which are crucial in promoting investment and growth. This paper finds an inverse relationship between savings rates and the informal sector when the informal sector is small. Once the informal sector reaches a certain size, further growth in the size of the informal sector boosts savings rates. The non-linear relationship is confirmed in both parametric and semi-parametric estimations. Rather than allowing the informal sector to grow unchecked, policy should focus on removing barriers for successful operation of business in the formal sector.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. Empirical studies show that the key drivers of informality are high taxes, a lack of social security and benefits, excessive regulation, culture, and poor governance (e.g. Loayza Citation1996; Johnson et al. Citation1997; Tanzi Citation1999; Botero et al. Citation2003; Dabla-Norris, Gradstein, and Inchauste Citation2008; Khamis Citation2012; Ordonez Citation2014).
2. To test these theories one needs to disaggregate income into its permanent and transitory components (e.g. Paxson Citation1992; Alderman Citation1996; Loayza and Shankar Citation2000; Vergara Citation2001; Andrade and Guillen Citation2014; Grigoli, Herman, and Schmidt-Hebbel Citation2014). The empirical results are inconclusive. Testing these theories is beyond the scope of this paper.
3. Research has shown that households may save more when facing periods of low income (e.g. Deaton Citation1991; Carroll Citation1998; Butelmann and Gallego Citation2000).
4. In essence, in countries with weak institutions (early stage of development) the choice is between no production and production in the informal sector, while in countries with strong institutions (late stage of development) the choice is between production in the formal sector and the informal sector. This helps to explain why the informal sector makes a positive (negative) contribution to output in countries with weak (strong) institutions.
5. Available at http://www.wider.unu.edu/wiid/wiid.htm.
6. Available at http://pwt.econ.upenn.edu/php_site/pwt_index.php.
7. Savings = −113.8559–0.1104informal +31.1805income −1.6066incomesq +35.1726growth
(0.69) (0.216) (0.048) (0.104) (0.466)
−1.3859remittances-1.5351corrupt +0.0358open −0.2973interest −0.2498inflation
(0.000) (0.128) (0.090) (0.019) (0.027)
R2 = 0.57 n = 122.
8. Results available upon request.
9. It is not possible to estimate a dynamic version of the model because of missing data and because informal sector data are only available for a short time period.
10. Results available upon request.
11. The U-shaped relationship between savings and income holds if GNP is used rather than GDP: Savings = −13.2349–0.7469 informal + 0.0082informalsq+10.8366income-0.5269incomesq
(0.489) (0.015) (0.029) (0.017) (0.055)
-0.0729dscp+0.0438open −0.2907interest
(0.001) (0.000) (0.001).
12. See Blundell and Duncan (Citation1998) for details and a helpful discussion of the implementation of this method.
13. Note that this also involves controlling for non-linear effects between Z and I.
14. It is worth noting that the confidence band proposed by Härdle (Citation1990) ignores possible approximation error bias. Correcting for this would complicate the expression considerably since the bias is a complicated function of the first and second derivatives of g(I). The bias tends to be highest at sudden peaks of the data and at the necessarily truncated left and right boundaries of the data. However, if h is chosen proportional to 1/n(1/5) times a sequence that tends slowly to zero then the bias vanishes asymptotically for the interior points; see Härdle (Citation1990) and Wand and Jones (Citation1995).
15. For the endpoints we chose the 1 and 99 per centiles of the distribution.
16. Alternatively we could start with a production function of the form Y = Kα(CAL)1-α (where C refers to rent seeking activities) and work through to the same steady state solution.
17. A similar equation is in Swaleheen (Citation2008) but no explanation is given as to its derivation.