A Call for Enterprise in
Economic Data Generation and Information Analytics
19th May 2017
Dr. Viral V Acharya
Deputy Governor, Reserve Bank of India
Presentation at the 9th Indian Chamber
of Commerce Banking Summit, Kolkata
State of Economic Research on India
A vibrant network is slowly but steadily emerging
University and business school professors
Analysts at banks, non-bank finance companies
(NBFCs), rating agencies, among others
Researchers at policy institutions and think tanks
Probing inquiries and fact discovery by media
Seminars, conferences, forums, panels, deputations
Global interest in studying India is surging
More undergraduate and post-graduate (MS, PhD)
students interested in pursuing Economics and Finance!
Miles to go before we sleep... on a good, firm trajectory
2
How Do We Accelerate?
The situation seems ripe for
Enterprise in
Economic Data Generation and Information Analytics
3
A HUGE opportunity!
Alongside banks and other financial intermediaries,
need a parallel ecosystem of economic and
financial data and information services that
Collects, collates and generates new data points
on the economy and financial markets
Disseminates publicly or sells the data
Analyzes, aggregates and researches data to
provide information analytics
Creates information-based business opportunities
Aids analysis-driven policy-making and thinking
Given our core human resource strength in
computing and information systems, this is a low-
hanging fruit that has not yet been plucked
4
Examples
Real-time inflation and consumption metrics:
E-commerce sites
What are the sustained temporal and geographic
variations in prices and quantities?
Employment statistics:
Payments data; bank and NBFC KYC data
Can Big Data help us compute quarterly unemployment
rate?
Rural and informal economy:
NBFC and Micro-finance institutions; FMCG companies
Do omissions of rural and informal economy in formal
statistics mask economically relevant growth and inflation
outcomes?
State finances:
Implied credit rating/risk using RBI State Finances report
What is the implied subsidy in borrowing costs?
5
Examples
Hot money flows:
Corporate bond, commercial paper, External
commercial borrowings, Masala bonds FPI
investments (maturity/location)
Which of the flows are "carry trades" and which are
long-term?
Governance and corporate finance of pyramids and
group companies:
Consolidate individual company/subsidiary filings
Are internal transfers tunneling or internal capital
markets in response to credit constraints?
Are foreign transactions round-tripping / tax-arbitrage
or genuine investments?
Bank lending boom and bust cycles:
Let me elaborate on this as a leading example with
one of my ongoing research studies and how it could
be done better
6
The Anatomy of a Business Cycle
Presentation at The 2nd Moody's, ICRA and NYU Stern Conference:
August 3rd , 2016
Viral Acharya Prachi Mishra N. R. Prabhala
New York University RBI CAFRAL, Univ of Maryland
Qualifier
Views are personal.
Not necessarily the official viewpoint of RBI.
8
Context
We analyze the anatomy of India's economic and
financial cycle since 2008
- Cycle is big
- Cycle is rather sharp
· Understanding and disentangling the channels
Bank lending channel
· Supply of credit too low?
· State-owned (distressed) banks
Corporate distress channel
· Demand for credit too low?
9
Overview: India's economic and financial cycle
Investment
Pick up in investment after GFC
Slowdown starting 2011-12
Similar cycle for other real outcomes
Similar cycle for bank credit
Credit and real cycles highly correlated
10
Real and Credit outcomes
11
Firm Sales and Employment
Growth (Annual average, in %) Capital Expenditures
Employment growth Sales growth (Firm-level, average, in %)
18 25
16 24
14 23
12 22
10 21
20
8
19
6
18
4
17
2
16
0
15
2008 2009 2010 2011 2012 2013 2014
-2 2008 2009 2010 2011 2012 2013
Notes. Capital expenditures (t) = (Net fixed assets (t+1) Net fixed assets (t)
+ Depreciation)/Net fixed assets
12
Growth in Credit: By Bank Ownership
(Annual, in %)
25 State-owned Private
23
21
19
17
15
13
11
9
7
5
2008 2009 2010 2011 2012 2013
13
% of Gross Advances Stressed Assets of Banks
Restructured Assets % Gross NPA % (RHS)
6.5 8.0
7.5
6.0 7.0
6.5
5.5 6.0
5.5
5.0 5.0
4.5
4.5 4.0
3.5
4.0 3.0
Jun-12
Sep-12
Dec-12
Mar-13
Jun-13
Sep-13
Dec-13
Mar-14
Jun-14
Sep-14
Dec-14
Mar-15
Jun-15
Sep-15
Dec-15
Mar-16
14
Credit and Investment Cycle
10 10
8
5
6
0
2008 2009 2010 2011 2012 2013 4
-5 2
0
-10
-2
-15
-4
-20 -6
credit growth (annual, in %) investment (right scale, in mn of Rs)
15
Can we disentangle
the bank lending (supply) channel
from
the corporate demand (demand) channel?
Should policy resolve bank stress or
corporate stress or both?
16
Empirical strategy: Diff-in-diff
· Do weak firms, and firms connected to weak banks,
respond differently from healthier firms, connected
to the same banks, when the cycle turned?
Weak and strong firms
Firms connected to weak or strong banks
Use variation pre and post 2012 when cycle
turned to distinguish bank lending channel from
corporate channel
17
Data
Firm-level real and financial outcomes
· CMIE Prowess
· 3,000 listed companies
Real outcomes
· Sales, employment, capx
Financial outcomes
· ICR, assets, leverage
Bank-level data
· BSR 2, Reserve Bank of India
18
Data (contd.)
Weak firm
Interest Coverage Ratio (ICR) < 2
Weak bank
Public sector banks
High Exposure to weak sector
Higher ex-post NPA
Firms connected to a weak bank
At least one bank is a PSB
Al least one bank has exposure to weak sector
(Max) non-performing assets: Above and below
median
19
Overview: channels
Bank lending channel helps understand the cycle
Firms connected to "weak" banks over-invested
and had better real outcomes in up-cycle, but with
much weaker outcomes during down-cycle
Firms with weak corporate balance sheets had
worse outcomes throughout the sample
Results provide a strong case for the asset quality
review and clean-up of banks underway in India
20
Employment growth by firm Employment growth by bank
stress stress
(weak-strong, in pp) (weak-strong, in pp)
0 1
2009 2010 2011 2012 2013 2014
0.5
-1
0
-2 2009 2010 2011 2012 2013 2014
-0.5
-3 -1
-1.5
-4
-2
-5
-2.5
-6 -3
21
Capx by firm stress Capx by bank stress
(weak-strong, in pp) (weak-strong, in pp)
0 2.5
2009 2010 2011 2012 2013
-2 2
1.5
-4
1
-6 0.5
-8 0
2009 2010 2011 2012 2013
-0.5
-10
-1
-12 -1.5
-14 -2
-2.5
-16
Notes. Capital expenditures (t) = (Net fixed assets (t+1) Net fixed assets (t)
+ Depreciation)/Net fixed assets
22
Interest coverage ratios by firm Interest coverage ratios by bank
stress stress
(weak-strong, in pp) (weak-strong, in pp)
0 0
-0.5 2009 2010 2011 2012 2013 2014 2008 2009 2010 2011 2012 2013
-1 -0.2
-1.5 -0.4
-2
-0.6
-2.5
-3 -0.8
-3.5
-1
-4
-4.5 -1.2
23
Empirical specification
Key Hypotheses: 1. Firms connected to weak banks
had poorer real outcomes once the cycle turned
Key Hypotheses: 2. Weaker firms had poorer real
outcomes through the cycle
24
Economic significance
Counterfactual exercise:
Losses from a firm's association with a weak bank
= How much higher would economic outcomes be if firms
were NOT associated with weak banks
(1) Overall change 2011-14 (2) Weak bank induced contraction
(% of 2011) (% of 2011)
(3) Real loss = (2)/[(1)+(2)] (in %)
25
Results
26
Economic significance
Counterfactual exercise:
Losses from a firm's association with a weak bank
Employment
(1) Overall change (2) Weak bank (3) Real loss =
2011-14 (% of 2011) induced contraction (2)/[(1)+(2)] (in %)
(% of 2011)
6.3 5.5 46.3
Sales
38.1 7.5 16.4
Capx
34.8 7.8 18.4
27
Conclusions from the Study
Bank lending channel important in explaining the cycle
Real outcomes stronger for firms connected to weak
banks in the up-cycle; but decline during down-cycle
Firms connected to weak banks have weak balance
sheets throughout the sample
lower ICR, higher leverage, are larger in size
Firms with weak corporate balance sheets had worse
outcomes throughout the sample
Results provide strong case for clean-up of stressed bank
balance-sheets by resolving heavily indebted firms
28
Corroborating Evidence
RBI Monetary Policy Report (MPR, April 2017) finds
supporting evidence using only bank-level data
Banks with greater stressed assets and worse capital
ratios / provision cover:
Lend at higher rates earning greater net interest
margins, but as a result
Show weaker credit growth
Bank-level analysis, however, makes it hard to rule out a
demand-based explanation that the bank became
stressed due to risky borrowers, which in turn are facing
higher rates and are not demanding credit any more
29
Questions Left Unanswered
Did healthier banks in a consortium lend more to
healthier firms compared to weaker banks?
Did stressed banks that responded with
recapitalization and provisioning lend healthily?
Did under-capitalized and under-provisioned banks
evergreen their bad loans lending to stressed
borrowers at over-subsidized rates to roll over debt?
30
Questions Left Unanswered
Did banks and firms that did restructure experience
better outcomes?
Did stressed banks have poor transmission of
accommodative monetary policy during 2015-16?
What did stressed banks do with excess liquidity during
demonetization compared to healthier banks?
31
Could we have done this better? YES!
1. Bank-firm loan-level matched data w/ loan terms at
time of origination and corporate finance data
- Should this be a public credit registry? Public good?
- All creditors, e.g., trade creditors also?
- E.g.: RBI BSR-RBI CRLIC-CMIE Prowess integration
2. Bank-firm loan-level ratings data
- Internal / external ratings and their evolution
- Market-based measures of firm and sector credit risks
3. Bank-firm loan-level restructuring data w/ details
- Augmented CRLIC
4. Platform for secondary loan sales and price discovery
5. Firm-debt level Default and Recovery (LGD) data
- Rating agencies should track and provide this
Such data could also help "lean against the wind" of a
lending cycle, e.g., with risk- and sector-based provisioning
32
Such datasets exist in many other countries
UNITED STATES, for example:
1. Deal Scan: syndicated loan origination
2. Shared National Credit Program: originations and draw
downs
3. Capital IQ: draw downs
4. FDIC Call Reports: bank statistics
5. SNL Financial: bank statistics
6. Dealogic: mergers and acquisitions
7. LSTA: secondary loan sales
8. Prowess/Losscalc: default and recovery rates
HMDA (mortgages), Survey of Small Business Finance, ...
33
Key Players
1. Large banks in commercial and mortgage lending, and
large NBFCs and micro-finance institutions in rural and
MSME lending can set data standards
2. RBI can play an aggregating role to collate data at
source from all financial firms and disseminate with
appropriate lags, if any
3. Data vendors and information analytics firms,
potentially housed as arms of large banks and rating
agencies, can distribute data and analysis
4. Vibrant research community I referred to at the outset
can be its consumer
5. Private financial firms can use analytics to undertake
analysis-aided enterprise and financial transactions
34
Summing Up
"Not everything that counts can be counted; and not
everything that can be counted counts."
- Albert Einstein
It is a sobering thought for economists!
It should induce innovations to count better what really
counts!!
Time ripe for taking giant strides in
Economic Data Generation and Information Analytics!!!
35
|