Ammu Lavanya on How Foreign Capital Changed Indian Bank Lending

SHRUTI RAJAGOPALAN: Welcome to Ideas of India, a podcast where we examine academic ideas that can propel India forward. My name is Shruti Rajagopalan, and this is the 2025 job market series, where I speak with young scholars entering the academic job market about their latest research on India. 

Our seventh and final scholar in the series is Ammu Lavanya, a PhD candidate in Economics at George Washington University. Her research is in the areas of International Finance, Monetary Economics, Empirical Banking and Financial History.

We spoke about her job market paper titled International Financial Flows, Credit Allocation and Productivity. We talked about financial liberalization in India, the 2004 banking reform that increased the ceiling on raising foreign equity its impact on  market value, lending capacity and increasing productivity through credit in India, the difference between private versus state owned banks, and much more. 

For a full transcript of this conversation, including helpful links of all the references mentioned, click the link in the show notes or visit mercatus.org/podcasts.

Hi, Ammu. It’s so nice to meet you. It’s so nice to finally have you here. I’m excited to talk to you. Suddenly, monetary economics is back in fashion. I’ve been reading a lot of monetary papers. In this particular paper, you look at a particular banking reform that happened in 2004, where India raised the ceiling on how much a domestic Indian bank could raise equity from foreign sources. It raised the limit from 49 percent to 74 percent.

Banks already had some foreign investors because the first ceiling was 49 percent. Now, because of this reform, you find that some banks were more exposed to the new ceiling. That is, they managed to receive more inflows because they could now go from 49 to 74 percent. The banks that saw their foreign shareholding rise also saw their market value increase, their borrowing capacity improve, and as a consequence of that, the total credit that they lent out also grew.

We can talk about it. You look at this at loan-level data for about 31 banks and 15,000 firms between 1998 and 2008. What you find, which is most interesting to me, these banks that were more exposed to the reform—not only did they actually lend more, they actually lent more productively. That is, they lent into more productive firms. You actually managed to put a number on it. This increase in productive lending, you estimate, led to an aggregate total factor productivity of Indian manufacturing by about 0.8 percent in the years after the reform, which is nontrivial.

First, is this a good way of summarizing what you have done and what you find? Then tell us more about how you got started on this.

Understanding Capital Inflows and Financial Liberalization

AMMU LAVANYA: Thank you for having me here. I think this is an excellent summary. What really started me off—I’m generally interested in looking at how capital inflows affect economic activity, especially in emerging markets, because these are mostly the recipients of capital inflows. In general, we all know about the 1991 liberalization reforms and how that put India on a different growth trajectory altogether.

A lot of the existing studies usually focus on the trade liberalization reforms and the delicensing reforms that happened in the early ’90s. Financial liberalization in India is a very gradual and slow process. In fact, it is still happening. What do I mean by financial liberalization? It’s basically capital inflows in the form of equity or debt flows from abroad to different domestic industrial sectors.

It’s been a gradual process. In fact, in this year’s budget itself, the foreign shareholding limit in the insurance sector was increased to 100 percent. I’m particularly interested in the financial sector liberalization, which is basically financial inflows to the banking or nonbanking sector. I wanted to focus on the banking sector because there has been a huge debate between the RBI and the government of India about whether or not to increase the foreign shareholding limits in Indian banks, especially private banks. There has been a lot of back-and-forth going on for the past two decades, in fact. This is what motivated me.

Concerns Around Foreign Capital and Hot Money Flows

RAJAGOPALAN: If I can interrupt you for a second, what is the concern in increasing these limits? Because I think that’ll shed a lot of light on what we find in your paper eventually.

LAVANYA: I think one is hot money flows. A lot of these inflows are in the form of institutional investment. Institutional investments can also be long-term. In my data, I do see that a lot of them do stay for a very long time, especially in good banks, but a lot of them tend to be short-term. Any monetary policy change in the US, and then the money just goes out. That leads to an increase in the risk of, say, balance sheet mismatches. Suppose if it’s a debt flow, then if banks have borrowed in foreign currency, then it leads to balance sheet mismatches.

One needs to be really careful about these things because they can lead to crises. I think the Asian financial crisis—the main cause of this was hot money flows that came into banks and firms in these countries. I think that’s where the RBI is good. It’s good to be careful, but at the same time, I think there’s too much restriction. I think an overall blanket ban is probably not the right solution. You can look at other things, such as hedging or other kinds of regulation, but I think, generally, having capital inflows is good because it improves the financial development levels of countries. In a country like India, you need as much investment . . .

RAJAGOPALAN: We’re starved of it. We’re starved of investment, right, in some sense?

LAVANYA: Yes. You need as much investment as you can get. Of course, you have to be careful. That’s always there.

RAJAGOPALAN: The third point? Sorry, you told me there are three points. You said the first two when I interrupted you.

LAVANYA: Oh. The third point was basically that a lot of the studies that look at financial liberalization focus on direct inflows to firms. A firm can borrow from abroad or can issue equity, which foreign investors can invest in. How does it affect productivity? This is only possible for big firms who are listed, and that’s a very small proportion of Indian firms generally. Most firms, even bigger firms, borrow from banks. I wanted to look at what happens when banks receive capital inflows in the form of equity. Also, I show later that this also leads to an increase in debt flows, but in the form of equity.

How do they intermediate these funds? Especially because we’re such a bank-dependent economy. Also, banks play a very big role in intermediating capital flows from abroad. I wanted to trace out what I refer to as the bank-lending channel of capital flows, as to what happens when banks are liberalized, what happens to their balance sheets, what happens to their lending, who are they lending to, and how do I quantify the effects? This is what I do in the paper.

The Banking Reform and Ownership Landscape in India

RAJAGOPALAN: It’s super interesting, but before we get into the actual reform, it would be helpful to know how bank ownership operated in India before the reform. For the most part, India has had two kinds of banks: domestic banks which are state-owned and domestic banks which are privately owned. Of course, there’s some back-and-forth in this because privately owned banks were nationalized overnight in 1969. Some state-owned banks have been merged with each other. All sorts of musical chairs have gone on, depending on which decade you pick.

Then there is a question of foreign-owned banks setting up an Indian entity which is entirely Indian. Then there’s a question of Indian private-owned banks and Indian state-owned banks, which can now raise some foreign capital in the form of equity. These are all slightly different things. Your reform looks at the last bit of it. Can you walk us through how bank ownership has operated in India through the decades?

LAVANYA: Of course, there was, we all know, the bank nationalization episode where, overnight, private-sector banks were turned into state-owned banks. I think I would like to start from the ’90s. In the early 1990s, because of the bank nationalization reforms, most of the banking sector was state-owned. There were some private banks still in existence. These are often referred to as the “old” private banks. You can think of Karur Vysya or Federal Bank. These were still in existence. They weren’t really government-owned. Then, in 1993, I think as part of becoming part of the WTO and everything, basically India allowed new private-sector banks to be set up.

This is when a lot of the private-sector banks that we see today—that’s when they were set up. Also, they allowed entry of foreign banks. Financial sector liberalization can be thought of as either the entry of foreign banks, where, say, Citibank sets up shop in India, opens some branches. People have looked at the impact of foreign bank entry. It can also be looked at as capital inflows to domestic banks, where domestic banks receive foreign capital and then intermediate that and work as intermediaries to onlend these funds.

This happens. Then over the ’90s, this is what the ownership situation was like. Another thing to keep in mind is, when we look at public- and private-sector banks during this era, the foreign shareholding limits in these banks were restricted. Out of the total equity, only 20 percent of the shareholding could be held by foreigners for both public state-owned banks and for private banks. Again, I think domestic investors are also limited in the kind of investments that they could make. This carried on till 2001.

In 2001, there was an attempt to have a reform where Indian private-sector banks, the foreign shareholding limits, they said they would increase it from 20 percent to 49 percent of the bank’s equity. It’s very interesting to look at what happens. There was a lot of policy confusion because one document says only FDI [foreign direct investment] shareholdings can be increased. Then the investors don’t know if it just refers to FDI or if it’s total foreign shareholding. Is it inclusive of FII [foreign institutional investment]? I think there was a lot of back-and-forth. I actually went through the newspaper reports, the budget documents, and everything.

RAJAGOPALAN: The notification after notification from the RBI.

LAVANYA: You cannot make sense of what is happening. I’m sure that’s what the investors also feel. I’ve seen the data as well. This reform didn’t really happen. In fact, I think there was just one bank that saw some increase in its foreign shareholding. It was ICICI Bank, just some increase, but most banks it still remained below 20 percent. Initially, I actually started off wanting to look at this one, this reform, and I saw that it had no impact at all. That was because there was just so much confusion with respect to what was happening.

Then, in 2004, again, a similar notification, but this time it was clear. It was basically wanting to increase the foreign shareholding limits in domestic private banks from 49 percent, even though it was confusing earlier, but it says 49 percent to 74 percent. It was for the entire foreign shareholding. It included all kinds of flows, be it foreign direct investment, foreign institutional investment, depository receipts in stock exchanges abroad, or even holdings by NRIs [nonresident Indians] or some strategic partnerships with other banks.

This is the reform that I focus on. Public-sector banks still, even till today, are capped at 20 percent. Their foreign shareholding is capped at 20 percent. This is what the situation is still like today. One thing that I’ve noticed is that a lot of new banks that are set up today have a lot of foreign investment as they start out. I think that’s what is also driving the debates about whether foreign shareholding should be completely liberalized in private-sector banks. 

Banks Most Affected and Patterns of Foreign Investment

RAJAGOPALAN: Before we get into how this impacts the nonbanking sector, manufacturing and so on, when the reform happened, are you able to figure out who was able to benefit the most? That is, which banks actually benefit and receive more investment? Now, presumably, the way you’ve set up the paper, it must be banks that were already exposed to some amount of foreign capital and foreign investment, right?

LAVANYA: Yes.

RAJAGOPALAN: What do you find exactly in number of banks in terms of percentage points and so on?

LAVANYA: Some very interesting things here that I was not expecting. Of course, banks that were closer to the limit, so closer to the 20 percent limit or right above because there was still some confusion—they see a huge increase. Some of the bigger banks see an increase up to 60, 65, 70 percent. On average, across all banks, the liberalization increased foreign shareholding to around 40 percent. Of course, there are some banks that don’t see a lot of capital inflows and some banks that see a lot, so the average is a little in between.

Something that was very interesting was, public-sector banks—the reform does not target them—see an increase in their foreign shareholding. Of course, within the 20 percent limit, but these are banks where you have 1 percent foreign shareholding, and they suddenly see an increase to 15 or 16 percent, which is very, very substantial, especially because these banks are really big. These are usually some of the bigger and, I would say, well-functioning public-sector banks. They also issue depository receipts that are traded in the New York Stock Exchange.

I was not planning to include public-sector banks in my analysis, but this was like, “This is excellent variation. I think I should be including them.” I think that was very surprising. Of course, it also depends on bank size. I have to control for that when I do my analysis. It is definitely the bigger banks that benefit more, but smaller banks also benefit. I think it’s substantial.

For the private sector, it’s bigger banks like ICICI Bank, HDFC Bank. These are really good banks. They’re still good. They’re considered to be good. These are the banks that see the most capital inflows. ICICI Bank, I really think it’s demand from the investor’s side, where they’ve sometimes even breached the 74 percent limit. I’m sure there are penalties for that, but there’s a lot of demand. I think ING Vysya was a bank that also . . . because it was a partnership with a bank from the Netherlands, ING increased its stake in the bank. Because they see that this bank is in a partnership with a foreign bank, it also led to an increase in capital flows there.

I think some other banks. One bank that also had a lot of . . . it was merged later on. I think there was something that went wrong later on, but it was Centurion Bank, which also saw a huge amount of inflows. These are the ones that are completely driving the foreign capital inflows. They’re also very actively traded. I also look at the foreign institutional investment flows, not just the stock levels. I see the trading. They’re actually really actively traded. I think that’s what also drives an increase in their market value of equity.

For public-sector banks, State Bank of India is always at the limit. It’s always at around 18, 19 percent. I think some of the better public banks like Punjab National Bank, Bank of Baroda, and Canara Bank, these banks are almost above. They were really negligible before, around 2 or 3 percent, and then they see an increase to 16–17 percent of their shareholding that is held by foreigners. Even the smaller public-sector banks see an increase. They have 5 or 6 or 7 percent of their shareholding being held by foreigners. I’m not saying it’s a huge amount, but it is substantial, especially because the policy was not targeted towards them. Generally, I think it’s a nice indirect effect as well.

RAJAGOPALAN: In general, overall, state banks have been more stable. They’re like a bet that isn’t going to go wrong. On the other hand, they’re never going to give you great returns for exactly the same reason. The way I interpret what you saw happening with state banks is just foreign investors balancing their portfolios, right?

LAVANYA: Yes.

RAJAGOPALAN: Even though it’s not targeted towards them, most investors have a diversified, well-balanced portfolio. Within this, the way to do that would be to pick this new bank, which is growing really, really fast, like a Centurion or something like that, and then balance it out with your Bank of Baroda or Punjab National Bank, something very standard like that, is the way I would think about it. Is that the right intuition?

LAVANYA: I think so. There’s also this implicit guarantee that, “Oh, even if something goes wrong, the government’s going to bail them out.” You can never really go wrong. Of course, your returns might be lower, but generally, yes, I think you’re right about this.

Impact on Borrowing Capacity and Lending Behavior

RAJAGOPALAN: Now, two things happen because of the reform. Before I get into that, how much does it increase the ability of the recipient banks or the banks that are more exposed or benefit more from the reform to actually raise capital? Do you see a significant change in their ability to raise resources?

LAVANYA: I do. It’s very difficult to trace out exactly how this capital is flowing from an increase in equity and how it’s affecting the bank’s ability to raise more funds. I do see, for instance, a 10-percentage-point increase in bank exposure leads to more than 40 percent increase in overall borrowings. This is in comparison with less exposed banks and controlling for bank size. It’s not like it’s just borrowing more because they’re bigger. They definitely increase their borrowings, and they borrow at a much cheaper rate. The share of the borrowings that comes from abroad also goes up substantially. It does increase. I think we can go into the mechanisms later on.

RAJAGOPALAN: The mechanism must be a straightforward market mechanism. Earlier, there were fewer credible parties and investors that could vet these institutions. Now there are more parties and investors who can vet these institutions, and the market is able to price the risk better of lending to these institutions, right?

LAVANYA: Yes, exactly. Of course, overall, theoretically, when a country liberalizes, investors can diversify as premium goes down, and you have better pricing. That’s why market value goes up. It also holds when you look at individual firms or banks. That’s why you see an increase in the market value of equity and also an increase in the net worth. Basically, a higher market value of equity leads to a decline in their balance sheet constraints, which is why they can borrow more at cheaper rates. I also showed net interest margin goes up and deposit rates haven’t really changed during this time, which means that banks are borrowing at a much lower rate.

You can also look at this by looking at how their funding mix changes. The main source of funding for banks is their deposits. Basically, banks take in deposits, and that’s what they lend out. When banks are diversifying, basically, you call them noncore liabilities, basically, borrowing from abroad. These are noncore sources of funding. I show that banks see a reduction in their reliance on their deposit financing and an increase in their reliance on total borrowings that they also lend ahead.

RAJAGOPALAN: Now, there are a few things happening. As you pointed out, there’s a 47 percent rise in total borrowing. There is a big increase in the foreign share. Within that borrowing, there’s a 40 percent rise in total advances and so on. Now, two things happen due to this reform. These banks are able to lend out more credit. Second, there is a reallocation in the way they lend out the credit. I’ll start with the first one. How are they able to lend out more credit? What is the precise mechanism at play there?

LAVANYA: When we see, say, an overall increase in lending, we have to be very careful about thinking about credit demand and credit supply effects. Are they lending out more because they can access funds more cheaply? Are they lending out more because the economy is booming and firms demand more credit, so that’s why they’re lending out more to these firms? Usually, I think it’s difficult to tease these effects out, and that’s why you need loan-level data. I use loan-level data between banks and firms.

Basically, when you have loan-level data, you can control for firm credit demand. You assume that credit demand for the firm is fixed, and then you can see how the change in a bank’s exposure increases the credit supply to these firms. Basically, the earlier number, 47 percent, was an increase in overall lending. But controlling for credit demand, a 10 percent increase in exposure leads to an 11 percent increase in just credit supply. This is just completely coming from the bank’s side because they are able to access funding more cheaply.

Again, I think it’s simply, why are they lending more? Because they have more access to funds. They can raise funds more cheaply, and they would want to lend it out. That happens because, again, the net worth goes up and their market value goes up.

RAJAGOPALAN: This doesn’t change anything that you are saying so far, but overall, the banking reform was not the only reform done through the ’90s and the early 2000s, right? There are these other reforms happening in other sectors in the economy, which presumably were building a certain amount of momentum, right?

LAVANYA: Yes.

RAJAGOPALAN: The economy is also growing really fast. I know you’re able to control it at the firm level by matching the loan-level data, but overall, people seem to be buoyant and positive about the economy and are willing to lend it out. In times of great uncertainty, even if you manage to raise more resources, domestically or from abroad, you’re just not going to lend it out. If you had seen a reform, for instance, during COVID, it’s unlikely they would have lent it out in quite the same way as we see here. This is also like a snapshot of a particular time in the Indian economy, I think.

LAVANYA: Of course. For sure. I think things would drastically change if I’m looking post the financial crisis. Especially between the period 2008 to ’13, ’14, when there was just so much happening, especially in the banking sector. I think things would change quite a bit. We can discuss about this later. Generally, I really think this is a time when the economy was doing well. The global economy was doing well, so you had a lot of capital inflows to emerging markets. I think that is definitely one thing to keep in mind.

Productive Lending and Screening Mechanisms

RAJAGOPALAN: Now, the second thing that you find is, not only did they lend out more, the type of lending—or rather, not really the type of lending—the people they chose to give loans to also changed. This is, to me, the most fascinating aspect of your paper. They’re actually able to give loans out to more productive firms that just have a better track record. When I think about it from the point of view of a bank, the way I normally think is, “What would a banker do?” They just want to make sure that they lend into sectors where their money is safe and the money will come back safely.

Now, that could be more productive sectors of the economy or less productive sectors of the economy. On the margin, it just depends on what’s happening in those sectors because some sectors may be more backstopped or insured by the government, or something else might be going on in those sectors that make them safer bets. But in this case, not only are they better at lending out and recovering their money, they actually pick the more productive sectors of the economy, which is the most interesting thing. What exactly is going on in terms of the mechanism of how this happens? What is your intuition for that?

LAVANYA: When I started this analysis about looking at what kind of firms are banks lending to, I was expecting it to go either way, because when we look at the literature, you can think of this as some kind of a credit supply shock or credit boom. When you have credit booms, the literature has really documented that banks, because they have a lot of money in hand, they tend to be very relaxed about who they’re lending to. They make really bad decisions at that time.

RAJAGOPALAN: Yes, sometimes.

LAVANYA: Yes. That’s one channel. The other channel is, if your funding constraints have become loose, you have more money in hand. Do you use that money to lend better or make better decisions? Maybe you invest more in your loan screening, and that’s why it allows you to make better lending decisions in the sense that you lend to more productive firms.

I look at this in two ways. One is what is called the intensive margin, where a bank, prior to liberalization, already had a lending relationship with the firm. After this particular reform, what happens to its lending to these firms? I find that they increase their lending to the more productive firms. I also control for collateral, because I think when a bank decides to lend, it either looks at productivity or returns on that particular investment project, or it looks at the collateral that the particular firm can provide.

I do both. I look at both productivity and collateral, and the combination of both, but I think the productivity bit really stands out. Basically, banks continue lending to firms that are more productive at the intensive margin. Then, at the extensive margin, the results are even more exciting. The extensive margin means what is the likelihood of a bank starting a new credit relationship.

Managerial Practices and Governance Improvements

RAJAGOPALAN: Before we get into that, what do you think happened to the functioning of a bank because of this changing in the ceiling of foreign investors? Basically, is it better human capital? Are they mimicking foreign practices? Is the equity coming with better practices? Is it coming with the ability to pay your bank manager more, and therefore attract a better kind of bank manager? What is happening within a bank that is more exposed to the reform? Maybe this will help us contrast it with state-owned banks or banks that are not as exposed.

This is really fascinating because everyone wants to do this. Everyone wants every firm, every participant in the financial sector, to make better decisions, to make sure resources go to their most valued and productive use. It’s very interesting that this reform is able to do that. That’s the reason I’m asking the question.

LAVANYA: Again, going back to theory, when you have equity liberalization, agency costs are supposed to go down. Basically, managers make better decisions. Does it happen? A lot of FII literature looks at this by looking at human capital investments. I find that banks that are more exposed do invest more in the human capital; basically, staff expenses as a ratio of their total income, or how much are they employing after controlling for bank size again. They do invest more. It’s a substantial amount. I definitely think there is some kind of improvement in loan-level decision-making at the managerial level.

The other thing is, when a bank wants to attract foreign investors, it has to adhere to some international corporate governance standards. If you want more capital to come in, foreign investors are only going to invest if you are actually doing well, if you’re performing well. I think that’s another thing. I think just better disclosure norms, better governance. I think that’s another thing that’s at play here.

The third thing that I noticed is that, even a particular foreign institutional investor, if they hold a certain amount of shareholding, say about 2 percent or about 5 percent, they’re actually part of the bank boards, so they have representation on the bank boards, especially the bigger banks. I see this in ICICI Bank. I see this in HDFC Bank. Of course, ING Vysya is a partnership. A lot of these banks also have some kind of partnerships, so I definitely think that is also at play.

One is to attract foreign investment. They are part of the bank board, so they have a say in decision-making of the bank. The other thing is, a lot of these foreign institutional investments, these investments are undertaken—or there are some investment bankers, like these big—JPMorgan Chase. A lot of these investment bankers are also involved in deciding whether or not to invest in these banks. When you have such global players, I think definitely domestic banks themselves want to really make themselves more attractive.

I would like to cite one statistic from the study. Basically, for banks that are more exposed to liberalization, a 10-percentage-point increase in exposure leads to a 5 percent decline in the probability that a loan that they make is restructured in the future. I think that’s really nice. Basically, I think that really shows that they’re making better loans.

RAJAGOPALAN: Here, there are two things that you said in the second point you made which are quite interesting, which is, to even attract the foreign investment, they need to have better norms, better governance, and so on. Presumably, this whole dance that was taking place between the year 2000 and 2004 probably helped in that period because the banks that were better banks actually got familiar with these things because it’s not like a switch you can turn on and off.

There’s an anticipation that, “Oh, we think we’re going to be able to raise more foreign investment. Now, what do we need to do for that? We probably need better norms on these five or six different margins,” and they start strengthening those. Then those are the banks that are picked.

The second thing I find interesting about what you say is there were always banks that could have made better decisions and banks that made worse decisions, but there was clearly some resource constraint. Either it could be putting in a new, better technology system or a data analysis system, or paying the managers better so that they can make better decisions or at least attract the human capital that makes better decisions. Those things couldn’t be done until they could get this inflow or injection of investment into those particular firms. This reform makes that very clearly possible, and then they are off to a different trajectory.

LAVANYA: Yes, for sure. I think just maybe this is an aside, but it also reminds me of the more recent UPI [unified payments interface] revolution, where bank managers can actually see transactions of the borrowers, and they can make better allocation decisions. Of course, it’s not the same reform, and there’s no foreign investors involved.

RAJAGOPALAN: It is. Exactly. The tech transfers, the knowledge transfers, transferring norms, practices, people sitting on boards, so many interesting things are going on. On the extensive margin that you were talking about, I want to go back to that. One of the things I noticed when I was reading your paper was, the more exposed banks are slightly more likely to start new relationships and less likely to end existing relationships. This is between the banks as lenders and their borrowers, right?

LAVANYA: Yes.

RAJAGOPALAN: Again, what is the mechanism at play? What exactly are managers being told by their leadership? Is it just, “Grow your loan book”? Do they have these KPIs, like, “You need to increase the book by so many percentage points”? Is it, “Diversify your loan book”? What are the conversations happening in these banks that’s leading to such specific changes?

LAVANYA: Again, it’s very interesting. At the extensive margin, there’s also a lot of heterogeneity. Before I look at the mechanisms of what’s driving them, I think, what is the likelihood of ending an existing relationship? Banks are less likely to end relationships with more productive firms.

RAJAGOPALAN: Of course.

LAVANYA: If they’re already lending to good firms, they’re not likely to end that. They have a positive probability of ending a relationship with less productive firms. They actually end. They don’t give them any more new loans. These are banks that are more exposed. Then, on the other hand, when we look at the likelihood of starting the new relationship, banks are more likely to lend—they’re more likely to lend generally overall, but the probability is higher for firms that are more productive.

The next thing that happens when you start a new relationship with a firm, the overall likelihood is greater that more exposed banks start new relationships. But conditional on starting a new relationship, the initial loan that they make is smaller than if you were a less exposed bank. I was not expecting this at all. I was like, “This is definitely better screening at play.” Over time, when you look at when the relationship deepens—so I just looked at the next three years—as the relationship deepens, they actually increase their lending to these firms that they give the smaller initial loan to.

That was something that really amazed me, that I was not expecting better screening to take place at this time. I just thought, “Oh, if we have more funds, we’ll just lend them out. Maybe we can check whether they’re more productive,” but you’re actually lending out smaller loans. You’re waiting and watching to see whether these firms are more productive, and then you only increase your lending to these firms that are more productive.

I definitely, again, think that this is—I think they’re just making better lending decisions. They’re screening better. I, again, think there’s something that managerial decision, loan level decision-making is at play. They are expanding. They lend more to all kinds of firms, but they’re definitely increasing their lending and screening better to better firms. They’re lending better.

Firm-Level Effects and Data Construction

RAJAGOPALAN: Now, what is happening at the firm level? How do you evaluate the marginal productivity of a single firm in this context? You’re looking at about 15,000 firms. 

LAVANYA: Marginal productivity—usually, I use this to classify firms into more productive and less productive. Generally, it’s very difficult to tell. Basically, this comes from your production function calculations. It’s very difficult to estimate a production function, but what is done in the literature, and what I also follow, is you look at the marginal revenue product of capital.

Since you can see revenues from firm balance sheet, usually, instead of estimating TFP [total factor productivity], you estimate total factor revenue productivity. From that, you can estimate marginal revenue product of capital. Given certain assumptions on the production function—you have a Cobb-Douglas production function, so some assumptions that you generally hold when you look at firm-level data—your marginal revenue product of capital is simply your firm’s revenue over firm’s capital. That’s an estimate of how productive a firm is.

If you look at all your firms and divide them using the median of that particular sector that the firm belongs to, you can divide them into more productive or less productive, depending on whether they’re above the median in that particular sector or not. I know it’s a lot of jargon, but simply, it’s just an indicator for firm productivity. You can calculate it using firm balance sheet variables.

RAJAGOPALAN: Now, you’ve separated these. These are high-productivity firms versus low-productivity firms. What you find is that in their relationship with these banks, after 2004, for high-productivity firms, their borrowing translates into higher investments in their total assets, fixed assets, and a higher wage bill. That is, they’re able to pay their employees more and presumably attract better human capital. Low-productivity firms tend to shrink their assets. In part, probably because bank credit has become . . . they’re able to access less of it, or the price of that credit has gone up, right?

LAVANYA: Yes.

RAJAGOPALAN: This is super interesting because now this is not the intended outcome of this particular reform, at least not directly. But because these banks are now better governed and making better decisions—and presumably, they don’t have all the state bank nonsense of “You have to lend X percentage of your portfolio into rural sector,” or some nonsense like that—they’re actually able to pick and choose these firms. If they pick better firms, the better firms make better decisions, which then automatically deprives less productive firms of oxygen. Is that the way to think about it?

LAVANYA: That’s perfect. I don’t think I could have explained it better. I look at a firm’s exposure to this liberalization.

Basically, I create a panel of firm-bank matched loan-level data. I get loan-level data from the Ministry of Corporate Affairs. This is publicly available data, but it’s actually a huge task to download the data. Why does the Ministry of Corporate Affairs have this? Every time a bank or any other lending institution lends to a firm . . . It’s all secured lending. We’re not talking about unsecured lending here. If it lends to a firm, it has to register that particular loan with the collateral, with the Ministry of Corporate Affairs. Otherwise, if it’s not registered, it’s treated as unsecured lending.

Basically, that results in all the banks registering all their secured loans to firms with the Ministry of Corporate Affairs. Then I have balance sheet data for firms from the CMIE Prowess database. For all the firms in the database, I basically go to the MCA [Ministry of Corporate Affairs] website and get their loans. It’s a massive task. I had to do this for 40,000 firms, and then I reduced. There are common identifiers. I can match bank balance sheet data from the Reserve Bank of India with the loans, with the firm balance sheet data from the Prowessdx. I get a panel of firm-bank matched data.

It was a massive task, but I really think it was worth it to just trace out the lending mechanisms from banks to firms. This is the database that I use. We already talked about how I look at bank-level exposures. How is a firm exposed to this particular banking sector reform? I know which banks a firm is borrowing from, and I create a firm exposure measure by using a weighted average of all the banks that the firm is borrowing from.

Basically, if firm A is borrowing from three banks, looking at the value of the loans that the firm is borrowing, I created a weighted average measure using the value of the loans as well as the exposure of the banks that it’s borrowing from. Then I, again, can classify firms as more exposed or less exposed using this. As you mentioned, firms that were more exposed to this liberalization, of course, saw an increase in the total amount of money that they can borrow from banks. This translates into higher investment. It translates into more wage bills, which you can take as—if they’re expanding—it’s more employment, or otherwise they’re just attracting better human capital.

Aggregate Effects and Decline in Misallocation

LAVANYA: Then overall, when we look at across firms, the dispersion of the marginal revenue product of capital falls. This is a very important concept in the economic growth and productivity literature, where a decline in the dispersion—which is simply the variance of the marginal revenue product of capital—if it’s falling, it’s indicative of basically a decline in misallocation, which indicates an improvement in overall productivity.

I was really excited to get this data and just translate it into actual, real effects, looking at firm balance sheet data. I was able to show that it leads to, through the bank lending channel—of course, there are many other things going on—but this particular bank lending channel does lead to a decline in misallocation.

RAJAGOPALAN: The matching you’ve done, I can’t even imagine how tedious that was, but it’s well worth the effort. Look, there are thousands of firms; there are hundreds of banks. But every firm is not borrowing from hundreds of banks. They’re borrowing from three or four banks. Every bank is lending to thousands of firms. This matching actually can very clearly pinpoint who is getting exposed to better loan managers in some sense, quite directly. That’s really what you’re looking at and how that helps all the downstream consequences.

It could be two things. Either it could be because of selection bias—these banks are just better at selecting—or it could be that once a bank like this has given you money, it just improves other factors, such as the bank’s supervision is much better. Their advice to you is much better. They’re better at structuring a loan for you. If this bank has given you the money, then maybe your relatives are more willing to give you money. Other people around you are more willing to give you money. It could be selection. It could be this other effect of being blessed by a good bank.

LAVANYA: A signaling effect.

RAJAGOPALAN: Exactly, a signaling effect. It’s incredible that figuring out who is the productive firm in the economy is neither easy nor obvious, yet the market process is able to do this just so well. 

LAVANYA: I was really excited, especially with the extensive margin results with the screening, waiting, and watching. I was like, “Oh, wow.”

RAJAGOPALAN: Did you expect to find this? Because why would someone do such a tedious task if they didn’t expect to find this? “How do you write a paper like this?” is my question, I guess.

LAVANYA: I was expecting to find an increase in overall lending for sure. I knew that balance sheets would be affected. A priori, I was expecting it to go either way. I had a story prepared that, “Oh, if they’re lending to less productive firms, it’s because it’s a typical credit boom to go where they’re just lending to everyone.” Generally, when you look at India’s banking sector, you expect such inefficiencies. I was really expecting that.

This really surprised me. There are two or three checks to see, and they match up. Even at the firm level, misallocation goes down. At the loan-level data, credit allocation is improving. I also check it at the sectoral level. Overall, again, at the sectoral level, variance of MRPK (marginal revenue product of capital) again goes down. It checks out. I’m pretty sure that if I had looked at a longer time period, maybe results might have been different. I don’t know. I think that’s something I want to do. Generally, given this time period, given all of this, I think I was really surprised and happy about it.

RAJAGOPALAN: I’m thrilled to find this. It confirms all my intuition about economic reforms, economic growth, and things like that. I’m totally on this side. One last part about your paper, which we didn’t get into, which I think is really important, is now you’ve figured out a way to map this firm-level productivity which you’ve carefully measured, and then figured out a way to aggregate it and think about the total factor productivity increase within one particular sector, which is manufacturing.

LAVANYA: I think I would like to think of it as a thought experiment. We know that credit allocation has improved. The mechanism that I use is taken from the literature. It’s a paper by Sraer and Thesmar that was published two years ago. Basically, the experiment here controls for general equilibrium effects. Assume that this is the only reform taking place in the economy. Nothing else is going on. What happens if this reform is scaled up?

It’s very difficult to think about scaling up in my context because it’s a continuous exposure measure. But I would think of it as, what if all the banks in the economy became more exposed, say, above the median of the exposure? If all the banks became more exposed and this was the only reform taking place in the economy, then through the bank lending channel, the aggregate TFP—it’s not just the manufacturing—the aggregate TFP increases by 0.8 percent.

I’m not saying this led to an increase of 0.8 percent. Of course, I think it led to a positive impact. Given these caveats that this is the only thing taking place, and if it was scaled up, the aggregate TFP . . .

Implications for Policy and the Future of Liberalization

RAJAGOPALAN: This is not trivial. That’s pretty good.

LAVANYA: That is not trivial. Comparing it to other particular reforms, there have been studies that have looked at different sectoral FDI liberalization. They find that direct equity flows to firms. For them, they find the lower bound is 3 percent, but then that’s direct flows to them, not to the bank lending channel. Anywhere between 3 to 16 percent, which is also not trivial, but just focusing on one particular channel through the bank lending channel, I find that the impact is 0.8 percent. I definitely think it’s not trivial. I think it’s quite a significant impact.

RAJAGOPALAN: That’s huge. I will take a 0.8 percent increase in TFP anytime. This is the dream situation. A couple of questions. Now, what is a policy implication of this in present-day India? Right now, the reform’s been done. We understand that all the crazy fears we had about exposure to foreign capital and hot money, all that has largely been worked out pretty well, both because of the caution shown within India by regulators, but also more generally. The Indian economy has been able to absorb that capital very, very well.

We also know that since 2004, various banks have come and gone, but also the banks that are left behind—there is an increase in their exposure to this kind of foreign ownership. Now, what is the present-day implication of that? Would you find the same effect if we went from 74 percent to 100 percent, if we just allowed a foreign bank to come set up in India and have its own loan book? What is the next step in this reform, if there were a next step at all?

LAVANYA: I think there’s a huge debate going on about whether you should increase it from 74 percent to 100 percent. I would say there would be improvements in productivity. But one always, again, has to be careful of hot money flows. I feel like the solution for that is not having overall restrictions on just capital. I think if you’re scared of hot money flows, if you’re scared of foreign currency exposures, I think regulation should target that thing particularly.

If you’re scared of banks being unhedged, make sure that they’re adhering to hedging norms. Make sure that the banks have a good capital base, that the quality of capital is good, so that even if they have capital outflows—suppose the US increases its interest rates and capital goes out—despite that, even if there are hot money outflows, the banks are not in a bad position, because all your other institutional checks are in place.

I think regulation should focus on that, make sure that the banks are well managed and they’re adhering to these capital quality norms, rather than imposing an overall blanket ban on capital inflows. I was reading very recently that there was a recent proposal of going back to 50 percent. Then the government rejected it.

RAJAGOPALAN: Who floated this proposal?

LAVANYA: I think the RBI is very conservative. I think it’s really, really worried about extensive capital outflows and losing reserves. If it was in the hands of the government, I do think it would already have increased to 100 percent, but the RBI is quite conservative about this. I can see why they’re conservative.

Differences Between Public and Private Banks

RAJAGOPALAN: A second question here. Everything we’ve talked about so far, do you see a clear difference between privately owned banks and state-owned banks?

LAVANYA: One thing I noticed: no causal impact. I haven’t done the regressions. Generally, a lot of the better public banks, public banks that receive foreign shareholding and had consistently higher levels of foreign shareholding—say between 15 and 20 percent—are still the better-performing banks today.

These are Bank of Baroda, Punjab National Bank, Canara Bank, and State Bank of India.

RAJAGOPALAN: Of course.

LAVANYA: Some of the banks which always had lower levels—they’re all merged now. They had these huge NPL [nonperforming loan] issues. Of course, the NPL issues were again due to a different regulation later on, but they had all these NPL issues. All of them are merged, like UCO Bank, Andhra Bank, and Vijaya Bank. 

RAJAGOPALAN: The NPL issue might be because of a different regulation, but the underlying cause of nonperforming assets is the same that you were talking about, which is a better loan book or a worse loan book. From everything you’ve told me so far, increasing foreign ownership makes your loan book far better.

LAVANYA: Yes, for sure. The NPL issue happened because of the regulatory forbearance reform during the 2008 crisis, and there’s been excellent research, I think, by a lot of Indian researchers on how it led to an increase in zombie lending. Basically, banks did not want to show these bad loans in their books, so they continued to lend.

RAJAGOPALAN: They just keep evergreening the same loan.

LAVANYA: Yes, evergreening. These are definitely the banks that did not have a lot of foreign ownership. These are definitely . . .

RAJAGOPALAN: Yes, because those banks would have people on the board who are asking you, “Why aren’t you marking this nonsense down?”

LAVANYA: Yes. I do think these banks eventually had to be merged or had to be bailed out. I definitely think wanting to attract foreign investors and having some checks and balances with independent directors and representation on the board also makes a big difference.

RAJAGOPALAN: I think it’s also a question of which battle you’re fighting if you’re the RBI at any given point in time. If your battle is, we need to increase the amount of investment; we need to have some catalyzing factors that will increase economic growth and lower NPLs—which are all, by the way, problems that the Indian economy is currently facing—then this seems to be the right thing to do. If the demons that you’re fighting are capital outflows and, “Oh my God, this is going to be the 1997 East Asian crisis all over again,” then you go back, you turn the clock back.

LAVANYA: Even if that is an issue, I think there are other things that . . .

RAJAGOPALAN: There are other ways to solve for that. I understand. It shouldn’t be done through ownership and equity stake. This was an absolutely fascinating read. I know you measure this till 2008 because then there’s the financial crisis, which is this whole other thing. What do you think you would find if you look at it from, say, 2013 to the present? Do you think the results would roughly hold, or at least 2013 to 2019, till COVID?

LAVANYA: From 2013 onwards, one thing is there was just a huge NPL crisis, especially for public-sector banks. That is between 2008 to 2013.

RAJAGOPALAN: That should only make your results stronger, not weaker.

LAVANYA: Yes, that is true, especially because it mostly affected public-sector banks. That’s true. Then I think controlling for a lot of things becomes a lot of trouble. Generally, from 2013, 2014 onwards, there have been a lot of new banks that have been set up.

RAJAGOPALAN: That’s true.

LAVANYA: When I look at the data post-2013, post-2014, a lot of them are just close to the threshold. Foreign shareholding is about 65 to 74 percent. I’m sure they just want the restriction to be removed. I think generally the RBI also tightened its regulating and NPL standards. I definitely think there would be an improvement in productivity. Lending would have gone up by a lot because there are so many new banks, so many banks that have a lot of these capital inflows that are persistent. It’s not like it’s just a short-term thing.

Of course, I can’t use this particular reform there, so I don’t have a clean experiment. I have to think of how I go about . . .

RAJAGOPALAN: Other ways to measure the same thing.

LAVANYA: Yes, like how to look at causal impacts there. I have the data. I just have to think of this. I really think the positive effect would hold up.

RAJAGOPALAN: I think so, too. This was absolutely fascinating. Very fun for me to read. Thank you so much for coming and talking about it.

LAVANYA: Thank you so much. This was a pleasure. I really had a lot of fun talking about it. Thank you.