b regress the individual indicators βb on log deposits of the bank in 1980 (sizeb), an indicator
variable (Natb) ,which takes the value of one when the size was larger than the cutoff and the bank therefore nationalized, and an interaction term (Natb ∗ sizeb) . This specification thus allows for a break at the nationalization cutoff value, as well as differential slopes for banks below and above the cutoff:
b βi = α + δ1sizei + γ1Nati + δ2 (Natb ∗ sizeb) + ε
[FIGURE 1 ABOUT HERE] [FIGURE 2 ABOUT HERE] Figure 1 presents the average share each bank provides to small-scale industry (top panel) and “trade, transport and finance” (bottom panel). In the figure, banks are ordered by the size of their deposits in 1980, so that banks below the cutoff of 14.5 are private, while banks
ˆ above were nationalized in 1980.21 The left line gives the relationship ˆα + δ1 ∗ sizei, while the
ˆˆ r i g h t l i n e g i v e s t h e r e l a t i o n s h i p ˆ α + ˆ γ + ³ δ 1 + δ 2
Contrary to the results obtained by
simple comparison of means, there does not appear to be any significant difference in lending to small-scale industry between public and private banks that are of similar size. That is, we cannot reject the hypothesis that nationalization had no effect on credit to small-scale industry. On the other hand, nationalization appears to have had the effect of lowering the amount of credit banks provide to trade, transport, and finance.
Nationalization appears to have had a large effect on credit to agriculture, as indicated by the top panel of Figure 2. There is a relationship between size in 1980 and lending to agriculture in 1992: larger banks lend more to agriculture. However, there is a visible break in the relationship at the nationalization cut-off: banks just above the cutoff lend substantially more to agriculture than banks just below, even after accounting for the effect of size. The analogous graph for rural credit is presented in the bottom panel of Figure 2.
[TABLE 4 ABOUT HERE] 21 To avoid disclosing bank-specific data, we have grouped banks with similar deposit size in 1980 into pairs or groups of three. Thus, while our sample includes 42 banks that were private or nationalized in 1980, there are only 19 points on the graph. The statistical analysis presented in Table 4 provides esimates based on individual bank-level data.