India's GDP Trends

So, I recently downloaded all this GDP and other macro data from the RBI’s database on the Indian economy (http://dbie.rbi.org.in), and I can hardly contain my excitement. My mind is like the inside of a fat water balloon – it is literally “bursting” with ideas for posts!

But since I’m an adult, I’m going to restrain myself and go step by step. I will do a series of posts based on this data. Today, I’m going to present and analyze trends in India’s Gross Domestic Product (GDP) all the way back from the 1960s.

GDP Trends  

I’ve created below a chart of economy-wide y-y GDP growth over the last 50+ years. The growth rates charted here are those of GDP at Factor Cost, at constant prices (04-05 base year). What does that mean?

• “GDP at factor cost” = GDP at market prices – Indirect taxes + Subsidies. It is a measure of GDP that excludes net indirect taxes paid to the government (these are taxes paid on good and services, and not on personal incomes). Since indirect taxes and subsidies are “transfer” payments (i.e. no output is created or economic value added in such transactions), it makes sense to exclude them while measuring GDP growth.

• “GDP at constant prices” is calculated using prices of products/services in a chosen base year (04-05 in this case), as opposed to GDP at current prices, which is calculated using prices of goods/services prevailing in the current year. GDP at constant prices allows us to measure “real” GDP growth i.e. growth due to expansion in output (and not just inflation) since prices are held constant.

Below are the key trends and analyses extracted from this data:

• Annual GDP growth rates of the 2000s three times more stable (less volatile) vs. those of the 1960s. 

The chart above shows that GDP growth was significantly more volatile in the '60s and '70s, than it has been in later years. GDP growth rates have become more stable (less volatile) and predictable over the decades. For the mathematically inclined (nerds like me), I’ve calculated below the arithmetic mean of y-y GDP growth rates in each decade, as well as their Standard Deviation (from the mean).

What is Standard Deviation you ask? Standard Deviation (σ or SD) is simply a measure of how spread out numbers in a population are around the population mean.

Look at the third column in the table above. It shows that in the 1960s, SD/Mean was = 93%, which means that on an average, yearly growth rates lay almost “1 mean distance” away from the population mean of 4%. In the ’90s, this distance was “0.34 mean”, while in the 2000s, it was “0.30 mean”. These numbers clearly indicate that growth has become much more stable. Infact, GDP growth rates in the 2000s were 3x more stable (less volatile) vs. the 1960s [SD/Mean of 2000s (30%) =1/3rd SD/Mean of 1960s (93%)].

• Why has GDP growth become more stable and predictable, you ask? It’s because Agriculture, the sector exhibiting the most volatility in y-y growth, now comprises just 14% of real GDP vs. 48% in 1960. Services, which exhibits the least volatility, now comprises 67% of real GDP vs. 37% in 1960.

There are three key sectors in our economy (any economy really) – Agriculture, Industry and Services. The Indian Agriculture sector has always been the most volatile in terms of GDP growth because it is heavily dependent on the monsoon for irrigation. 

In the table below, I’ve calculated the mean & SD of y-y sectoral GDP growth rates for the period 2000-01 to 2013-14. The SD for the agriculture sector is 1.3x the mean! Kind of defeats the purpose of calculating a mean, huh? Services growth is the most stable (SD = 0.22x mean).

Given that the agriculture sector is now (2013-14 estimates) just 14% of real GDP vs. 48% in 1960, and Services comprises 67% of GDP vs. 37% in 1960, GDP growth today is much more stable and predictable.

• Average y-y GDP growth has been accelerating over the decades (3.2% in the 1960s to 7.2% in the 2000s). The answer lies again in sectoral composition --> in the rising share (67% of GDP today vs. 37% in 1960) of the services sector where growth has been accelerating; and the falling share (14% today vs. 48% in 1960) of the volatile and low (avg.) growth agriculture sector. Industry has exhibited a modest though patchy/inconsistent acceleration in growth (19% of GDP today vs. 14% in 1960). See chart below and pie charts above.

• How does India’s GDP growth compare with other nations? For the period 2006-14, India’s GDP has grown at an average 7.5% y-y vs. just 1-1.5% growth for the major developed nations (USA, UK and Germany) and 3-3.5% for the emerging economies of Russia and Brazil. China has grown at a scorching 9.9%. That said, growth has slowed down in China this year and it faces many challenges as it looks ahead.

India though well positioned for now, has challenges of it’s own (subject for another post). I’m going to talk about Sectoral GDP trends in detail in our next post. Ciao for now!

Time To Narrow The Gender Chasm in National Income Accounting

Want a break from all the money market stuff (Repos, CRR etc.) I’ve been talking about so far? That makes the two of us.

I have just what the doctor ordered - some mid-day comic relief from one of my favorite webcomics.

This also happens to be a great segue to the relatively serious stuff I’m going to be talking about in this post, which as you may have guessed, has something to do with “gender chasms” and something to do with “national income accounting”.

Follow the link below to read my full article “Time To Narrow The Gender Chasm in National Income Accounting”, published by Point blank 7 in May 2014.

Time to narrow the Gender Chasm in National Income Accounting

Point Blank 7 is an upcoming, independent news website focused on presenting a wide range of news and opinions (sometimes ignored by mainstream media) to its readers.

In this article, I’ve talked about the non-inclusion of unpaid domestic services (in developing countries, 90% of these services are provided by women) in the United Nation’s globally followed “System of National Accounts” and its devastating consequences for women. It’s a simple, quick and unpretentious read – great for gaining a speedy grip on this important issue.

MY PROMISE TO YOU:
If you’re a woman (or man) who’s feeling under-appreciated for all the mind-numbing yet life-sustaining chores you do everyday for your loving yet slightly ungrateful family (its OK to say that out loud, even if its grammatically incorrect), this will strike a special chord in your heart.

Note: Incase you were wondering (I know you weren't, but I'm going to tell you anyway) where grammatical liberty was exercised above - it was in the phrase "slightly ungrateful". It's like saying "I'm slightly dead".

With love,

- The grammar Nazi

Calculating NDTL & CRR – in Practice

So in my last post titled “CRR: How to calculate Net Demand and Time Liabilities (NDTL) – the Theory”, I covered all the theory behind the NDTL calculation for CRR purposes. In this post, we’ll do a real-life numerical illustration in true Nerdverve “Ishstyle”.

Provided below is the same table that we included in the previous post - table number 4, “Scheduled Commercial Banks – Business in India” from the RBI’s WSS published on Sep 18th 2015. We’ll use the information in this table to calculate NDTL for SCBs (for CRR) outstanding as on September 4th.

From the table above (as on Sep 4^th):

Liabilities to Others = Rs. 97,341.5 Bn  (90, 280.5 + 2,380.2 + 4,680.8)

Liabilities to the Banking system = Rs. 1,806 Bn  (1,267.5 + 470.4 + 68.1)

Assets with the Banking system = Rs. 2583 Bn (1,772.1 + 207.7 + 232.8 + 370.4)

Since Liabilities to the Banking system are < Assets with the Banking system, NDTL = Liabilities to Others = Rs 97,341.5 Bn.  

********************************************************************************

Remember (from our previous post), NDTL for the banking system is =

Liabilities to Others in India (#2 in table above) + Liabilities to the Banking system (#1 in table above) – Assets with the Banking system (#5 in table above), when “Liabilities to the Banking system” – “Assets with the Banking system”> 0.

If “Liabilities to the Banking system” – “Assets with the Banking system” < 0, then NDTL = Liabilities to Others in India.

******************************************************************************

Now, from this NDTL figure obtained above, we need to subtract the following items that are exempted from CRR maintenance. (Please read my previous post for detailed explanations on all these exempted items)

1.    While calculating the “Net inter-bank liabilities” (“Liabilities to the Banking system” – “Assets with the Banking system”), SCBs should not include inter-bank term deposits / term borrowing liabilities of original maturities of 15 days and above and up to one year in “Liabilities to the Banking system”. Similarly, banks should exclude their inter-bank assets of term deposits and term lending of original maturity of 15 days and above and up to one year in “Assets with the Banking System”.

2.    Credit balances in Asian Clearing Union (US$) Accounts.

3.    Demand and Time Liabilities of banks’ Offshore Banking Units (OBU)

4.  The eligible amount of incremental FCNR (B) and NRE deposits of maturities of three years and above from the base date of July 26, 2013, and outstanding as on March 7, 2014, till their maturities/ pre-mature withdrawals.

5.  Minimum of Eligible Credit (EC) and outstanding Long Term Bonds (LB) to finance infrastructure loans and affordable housing loans, as per the circular DBOD.BP.BC.No.25/08.12.014 /2014-15 dated July 15, 2014.

Detailed data that would allow us to exclude the exempted items above from our NDTL calculation is not provided in the RBI’s WSS and other public releases. Banks however obviously possess all this information for their own operations. They subtract the exempted items from the NDTL calculated in the formula above, to get the precise NDTL for CRR maintenance purposes.

NDTL as on Sep 4th used to calculate CRR funds requirement for fortnight of Sep 19th – Oct 2nd 

Per RBI regulations, every alternate Friday, banks have to release certain data on their assets, liabilities, operations etc. These alternate Fridays are called “Reporting Fridays”. The period starting from the Saturday immediately following a Reporting Friday, all the way to the next reporting Friday is called “Reporting fortnight”. Let’s clarify with a real life example. The month of Sep 2015 had two reporting Fridays (see the calendar snapshot provided below) – the 4th and the 18th. The two reporting fortnights in the month were 5th -18th Sep and 19th Sep – 2nd Oct (this one included the first couple of days of October).

CRR maintenance happens with a fortnight’s lag. This means that CRR is maintained on the NDTL of the reporting Friday of the second preceding fortnight. For instance, for the reporting fortnight 19th Sep – 2nd Oct, CRR is maintained on the NDTL as on 4th Sep – the reporting Friday of the 2nd preceding fortnight (22nd Aug – 4th Sep).

This is why we are going to use the NDTL as on Sep 4th to calculate the funds requirement for CRR maintenance for the fortnight of 19th Sep -2nd Oct. 

Funds required for CRR maintenance during fortnight of 19th Sep -2nd Oct?

The NDTL we calculated above (Rs 97,341.5 Bn as on Sep 4th) is not adjusted for CRR exemptions. Lets use it however, to do a rough calculation of the CRR funds requirement for the fortnight of 19th Sep – 2nd Oct. CRR was 4%, hence the funds required for CRR maintenance were = 4% * 97,341.5 = Rs. 3,894 Bn. 

Note: The precise CRR funds requirement for a reporting fortnight is provided in the WSS released on the Friday following that fortnight. This supplement also includes the actual cash balances maintained by banks (with the RBI) during the fortnight. For instance, the CRR requirement for the fortnight of 19th Sep – 2nd Oct was provided in table no. 3 of the WSS released by the RBI on Oct 9nd (please see exhibit below). It was = Rs. 3,682.5 Bn. 

Our rough estimate of the cash reserve requirement was ~6% higher than the actual figure. Not bad I say! (given that we worked with incomplete data) 

We can work backward and calculate the actual NDTL figure (for CRR purposes) as on Sep 4th by dividing 3,682.5 by 4%, which gives us Rs. 92,062.5 Bn. 

How CRR is maintained in practice

We know from the data above that the daily CRR funds requirement for the fortnight of Sep 19th – Oct 2nd was Rs 3,682.5 Bn. This means that for the entire 14 day period, banks had to maintain total funds (adding up cash reserve reserves maintained on each of the 14 days) of Rs 3,682.5 Bn * 14 = Rs. 51,555 Bn. 

Per current guidelines, all banks are required to maintain a minimum CRR balance of 95% of the average daily reserves required for the fortnight, on all days of the fortnight. This means that banks were required to maintain at least Rs. 3,498.4 Bn (95% of 3,682.5) in cash reserves with the RBI every day of the fortnight in question, in addition to fulfilling the total funds requirement of Rs. 51,555 Bn for the fortnight. 

Lets assume that banks maintained the minimum balance of Rs. 3,498.4 Bn (95% of requirement) on the first 13 days of the fortnight. This means that on the last day of the fortnight, they would have to maintain a cash balance of (51,555 – 3,498.4 * 13) = Rs 6,076.1 Bn. 

In the table above, you can see that the total cash balances maintained with the RBI during the fortnight (adding all 14 days) were = Rs 52,605.4 Bn. This was Rs 1,050 Bn more than what was required for CRR maintenance over the 14 days.

This was the case because the banking system was in an excess liquidity situation. This fact is underscored by the tables provided below from the RBI's Weekly supplements (table no. 8 “Liquidity Operations of by the RBI) – they show the net absorption of liquidity by the RBI from the banking system through its Reverse Repo and Open Market Operations during the fortnight in question. Including the 19th of Sep (not shown in the tables below), the RBI absorbed a total of Rs. 224 Bn from the banking system during the fortnight.

In a tight liquidity situation, the cash balances maintained with the RBI would normally not have been in excess of the CRR requirement.