6.12 Case studies

6.12.1Rakesh Jhunjhunwala and Titan; his principles and his mistakes#

Context Rakesh Jhunjhunwala (1960–2022) entered the market in 1985: "In 1985, I was a Marwari chartered accountant entering the stock market" (Outlook Business interview, 3 Apr 2015). Figures for his starting capital and early profits circulate widely but trace to press profiles, not to him, and are not used here. He invested through Rare Enterprises, his proprietary firm, and died on 14 August 2022.

What is reliably documented about Titan The primary record is Titan's quarterly shareholding pattern, filed with BSE and NSE and posted on Titan's investor page (https://www.titancompany.in/index.php/investors/shareholding-pattern). Under the LODR format it names every public shareholder holding more than 1% (6.1).

• Early holding. Jhunjhunwala and his wife Rekha held 1.57% of Titan in June 2002, as reported by Fortune India from that quarter's shareholding pattern (secondary: the 2002 filing predates the exchanges' online archive; treat the figure as reported).
• Recent holding. The pattern for the quarter ended 31 Mar 2026 lists Rekha Jhunjhunwala with 4,71,84,470 shares, or 5.31% (as reported by Business Today, 12 May 2026, from the filing; check the current quarter's pattern on the exchange before quoting it). The family held through deep drawdowns, and Titan has long been its largest listed holding by value (holding × price, computable from the filing).
• In the same interview he recalls: "Titan's share price was ₹30 in October 2001" (Outlook Business, 3 Apr 2015). That is a market price, not his cost.
• Raamdeo Agrawal explained the thesis in a Business Today piece of 27 June 2026: backing Titan when few did.
• Entry prices are not reliably documented. Figures circulate on blogs, but none traces to a filing or to his own statement, and this course does not use them. Shareholding patterns name holders above 1% each quarter; they never disclose cost. Anyone quoting his buying price for Titan, Sesa Goa or CRISIL is quoting folklore.

His explanation of holding through the drawdown (Outlook Business interview, 3 Apr 2015): "I felt that there was no permanent impairment of value."

Mechanism In this module's terms the Titan thesis (as described by Agrawal) is a long competitive-advantage period: organised, trusted jewellery retail taking share from unorganised jewellers, with a brand moat that kept RONIC above WACC while the addressable market grew. The drawdowns were moves in price, not in the value-driver inputs. Jhunjhunwala's sentence is precisely the distinction between price volatility and impairment of NOPAT, $g$ or RONIC. The discipline it requires is to define, in advance, what would constitute impairment (a lost moat, a governance failure, a permanently lower RONIC) and to check it, rather than treating every drawdown as noise.

His principles in his own words (all from the Outlook Business interview, 3 Apr 2015, "You can't make money on borrowed knowledge"):

• "Tips are hazardous to your financial health and you can't make money on borrowed knowledge."
• "Trading is like a T20 match, you give and take quickly. Investing is like a test match."
• "You have to be dogmatic." And: "If you think you are always right and the market is wrong, you will not learn anything."
• On Lupin: "I was the second-largest investor and my conviction paid off."
• On losing positions: "I had blind faith in both the companies but I have learned from my mistakes."
• "People hope to double their money without realising that they could lose it too."

Lines such as "buy right, sit tight" are widely attributed to him but could not be traced to a primary source; treat them as attributed only.

His mistakes

• DHFL. Business Standard reported in April 2019 that he raised his DHFL stake in the March 2019 quarter (Q4 FY19). In January 2019 Cobrapost had alleged that about ₹31,000 cr was diverted through shell companies (an allegation at the time; later MCA, ED and CBI probes found promoter-linked entities; unverified). DHFL's commercial paper was downgraded to "D" (default) on 5 June 2019. The red flags of 6.8.3 and the NBFC ALM mismatch of 6.10 were visible at the time he added.
• A2Z Maintenance was described as his biggest dud; the company went into corporate debt restructuring (Outlook Business; the entry year conflicts across sources).
• Mandhana Retail fell about 94%; his estimated loss was about ₹40.7 cr (Zee Business; unverified).
• DB Realty is listed among his misses by Business Standard.
• Aptech insider-trading case. SEBI alleged that he and Utpal Sheth held unpublished price-sensitive information about Aptech (the UPSI period ran from 14 Mar to 7 Sep 2016) and communicated it to the other applicants, who traded on it. The seven applicants, including Rakesh and Rekha Jhunjhunwala, settled in July 2021 under the SEBI (Settlement Proceedings) Regulations, 2018, without admitting or denying the findings. For him alone the settlement order lists a settlement amount of ₹9,50,85,704 plus disgorgement of ₹5,86,79,870 and interest of ₹3,10,79,431 (SEBI settlement order SO/EFD-2/SD/384/JULY/2021, "Rakesh Jhunjhunwala and 6 others in the matter of Aptech Limited"). Press totals for all seven (₹35–37 cr) add the other applicants' amounts; sum the order's table yourself before quoting one.
• In a 2017 talk, asked about mistakes, he said he had made "a number of them" (Business Standard).

6.12.2Satyam Computer Services, January 2009#

Context Satyam was one of India's large IT services firms, in the Nifty and listed on the NYSE, and audited by a Big Four affiliate (Price Waterhouse).

Timeline and numbers

• December 2008: Satyam's board approved, then abandoned after investor protest, the acquisition of Maytas Properties Ltd and Maytas Infra Ltd, companies promoted by chairman B. Ramalinga Raju's family (SEBI order WTM/RKA/SRO/64-68/2014 of 15 Jul 2014, which records the "failed attempt"). Raju's own letter calls it "the last attempt to fill the fictitious assets with real ones". (A deal value of about US$1.6 bn was widely reported at the time; it is not in the SEBI order and is not used here.)
• 7 January 2009: Raju's letter to the board, e-mailed to SEBI the same day, confessed that the balance sheet of 30 September 2008 contained (₹ cr; the letter, quoted in full in paragraph 1 of the SEBI order of 15 Jul 2014, gives the figures in billions of rupees):
  • inflated (non-existent) cash and bank balances of 5,040 (against 5,361 reported), that is, 94% of reported cash was fictitious;
  • non-existent accrued interest of 376;
  • an understated liability of 1,230;
  • overstated debtors of 490 (against 2,651 reported).
  • Sum: $5{,}040 + 376 + 1{,}230 + 490 = 7{,}136$, the "₹7,136 cr" figure.
  • For the September 2008 quarter, revenue was reported as ₹2,700 cr against actual ₹2,112 cr, and operating margin as ₹649 cr (24%) against actual ₹61 cr (3%). The letter adds that this alone raised the artificial cash and bank balances by ₹588 cr in that quarter.
• The stock fell about 78% that day and the Sensex 7.3% (contemporaneous press reports; check the BSE and NSE bhavcopies of 7 Jan 2009 before quoting either to the decimal). Reports of the subsequent low differ.
• 11 January 2009: the government nominated Deepak Parekh, Kiran Karnik and C. Achuthan to the board. On 15 Jul 2014 SEBI's whole-time member barred Raju and four other former officers from the securities market for 14 years and ordered disgorgement of their wrongful gains with 12% simple interest from 7 Jan 2009 (SEBI press release PR No. 67/2014; the amounts were re-determined in later orders after appeals). A special court convicted Raju and others in April 2015 (reports give 9 or 10 April). The US SEC fined five PwC India affiliates US$6 m (SEC press release 2011-82, 5 Apr 2011). SEBI later barred the Price Waterhouse network firms from auditing listed companies for two years (SEBI order of January 2018; verify the date on sebi.gov.in).
• Raju later denied writing the letter, and his defence called it a hoax. The figures above are what the letter states, and the SEBI order relies on them.

Mechanism Every number in the confession is a violation of an identity in 6.2. Fake revenue with fake receivables would have shown as a low CFO/PAT ratio, so the receivables were "collected" into fake bank balances. That kept CFO looking healthy but created the second anomaly: a cash pile of over ₹5,000 cr that (a) earned interest that had to be fabricated too (₹376 cr of non-existent accrued interest), and (b) was never returned to shareholders through buybacks or large dividends, while a profitable IT firm was not using it. The checks that would have caught it:

1. Implied yield on cash. Interest received in cash divided by average cash, compared with deposit rates. Fabricated accrued interest inflates interest income but not interest received; the gap between accrued and received is the tell.
2. The capital-allocation puzzle. A business holding cash equal to a large share of its market value, not paying it out, and not buying back stock, needs an explanation.
3. Bank confirmations. An investor cannot do this; an auditor can and must. The failure was the auditor's.
4. Related-party signal. A plan to spend company cash on promoter-family property businesses is the strongest governance red flag in 6.8.3.

6.12.3Yes Bank: NPA divergence, 2016–2020#

Context Yes Bank grew its loan book far faster than the banking system through the mid-2010s, with concentrated lending to stressed corporate groups.

Timeline

• Why the divergence became public. RBI circular DBR.BP.BC.No.63/21.04.018/2016-17 of 18 Apr 2017 requires a bank to disclose, in the notes to its accounts, a reconciliation of reported against RBI-assessed gross NPAs, net NPAs, provisions and profit when the divergence crosses set thresholds; the first disclosures (for FY16 divergences) were due in the FY17 accounts. SEBI circular CIR/CFD/CMD/80/2017 of 18 Jul 2017 added immediate disclosure to the exchanges for listed banks.
• FY16: the RBI's annual supervisory assessment found gross NPAs of ₹4,925.7 cr against ₹749 cr reported: a divergence of ₹4,177 cr, about 5.6 times the reported figure (disclosed in Yes Bank's FY17 annual report under the 2017 circular; figures as reported by Business Standard, May 2017).
• FY17: RBI-assessed GNPA ₹8,373.8 cr against ₹2,018.6 cr reported: divergence ₹6,355 cr, the RBI figure about 4.1 times the reported one (Yes Bank's disclosure to the exchanges with its September 2017 quarter results; figures as reported by Business Standard, 27 Oct 2017; check the filing on BSE/NSE).
• September 2018: the RBI curtailed CEO Rana Kapoor's tenure (Mint, 24 Sep 2018: "Yes Bank board to meet Tuesday after RBI directive on Rana Kapoor's tenure").
• February 2019: the bank issued a press release saying the RBI had found nil divergence for FY18. The RBI publicly rebuked it for disclosing a confidential supervisory report and threatened action (Business Standard, 15 Feb 2019).
• 5 March 2020: the central government placed the bank under a moratorium to 3 April 2020 (RBI press release 2019-2020/2022, "Yes Bank Ltd. placed under Moratorium", which cites a "steady decline" in the bank's financial position and the absence of "a credible revival plan"). The government's moratorium order capped withdrawals at ₹50,000 per depositor. The same day the RBI superseded the board for 30 days under section 36ACA of the Banking Regulation Act, 1949, "owing to serious deterioration in the financial position of the Bank", and appointed Prashant Kumar, former DMD and CFO of SBI, as administrator (press release 2019-2020/2025). On 6 March ICRA downgraded the bank's bonds to "D" (press reports).
• 6–13 March 2020: the RBI published a draft reconstruction scheme (press release 2019-2020/2028, 6 Mar 2020) with SBI as the anchor investor; the government notified the Yes Bank Ltd. Reconstruction Scheme, 2020 on 13 March (RBI press release 2019-2020/2091), and the moratorium was lifted on 18 March 2020. SBI and several private banks subscribed new equity under the scheme; press totals differ, and the investors' own exchange filings give the individual amounts.
• AT1 bonds of ₹8,415 cr were written down under the scheme in 2020, and the write-down was litigated. Press reports state that on 15 October 2026 the Supreme Court upheld the quashing of the write-down and that the shares fell about 20% intraday (secondary; read the judgment on sci.gov.in before relying on it).

Mechanism For a bank, book value $B$ is loans minus provisions. Under-recognising NPAs overstates $B$ and current ROE. Through (6.18), the market priced Yes Bank at a premium P/B on a ROE that included income on loans that were already bad. A divergence of five times reported GNPA means B and ROE were both wrong. The repeated divergence (two consecutive years), the regulator cutting the CEO's tenure, and the bank publicising a confidential supervisory finding were governance signals on top of the accounting one. Loan growth far above the system with concentrated exposure is the bank analogue of revenue growing faster than cash collection.

6.12.4DHFL: a listed lender's funding and governance collapse, 2018–2021#

Context Dewan Housing Finance Corporation (DHFL) was a large listed housing-finance company, an F&O stock, and one of the most widely held NBFCs by mutual funds. It is the listed Indian case that combines the two lender failure modes of this module: the asset-liability mismatch of 6.10 and the governance and related-party flags of 6.8.3.

Timeline (primary sources where available)

• September 2018 onward: after the IL&FS defaults (6.10), short-term funding for NBFCs froze. DHFL had funded long-tenor home and project loans partly with commercial paper and mutual-fund money, and its shares fell sharply in late September 2018 as its paper was sold at distressed yields (press reports; check the NSE bhavcopies before quoting a percentage).
• January 2019: Cobrapost alleged that about ₹31,000 cr had been diverted through promoter-linked shell companies. DHFL denied it. Later agency probes are reported but are not used here as findings.
• 5 June 2019: rating agencies downgraded DHFL's commercial paper to "D" (default).
• 20 November 2019: the RBI superseded DHFL's board under section 45-IE(1) of the RBI Act, 1934, "owing to governance concerns and defaults by DHFL in meeting various payment obligations", and appointed R. Subramaniakumar, ex-MD and CEO of Indian Overseas Bank, as administrator. It announced that it would start resolution under the insolvency rules for financial service providers (RBI press release 2019-2020/1230). An advisory committee followed on 22 November (press release 2019-2020/1246).
• 3 December 2019: the NCLT, Mumbai, admitted the RBI's application. DHFL was the first financial service provider resolved under the Insolvency and Bankruptcy (Insolvency and Liquidation Proceedings of Financial Service Providers and Application to Adjudicating Authority) Rules, 2019.
• 7 June 2021: the NCLT approved Piramal group's resolution plan, reported at ₹37,250 cr including about ₹12,700 cr of upfront cash, against admitted claims far larger (press reports of the order; read the order for the claim figures). Equity holders received nothing under the plan and the shares were delisted (press reports; the NCLT order and DHFL's final exchange filings are the primary record).

Mechanism Two failures compounded each other.

1. Liquidity. A lender funding assets of duration $D_A$ with liabilities of duration $D_L\ll D_A$ must refinance a fraction of its book every quarter. When the refinancing market closes, it must sell assets or default, whatever the loans are worth. The ALM statement in the annual report (maturity buckets of assets and liabilities) shows the gap; the share of commercial paper in borrowings measures how fast it bites.
2. Asset quality and governance. Large wholesale loans to builders and entities that are hard to trace, plus public allegations of diversion, mean the book value $B$ in (6.18) may be overstated, as with Yes Bank (6.12.3). P/B is then meaningless.

A private-company postscript: Byju's and the auditor-exit signal Byju's (Think & Learn Pvt Ltd) is not listed, so a listed investor never held it, but the episode is the clearest recent Indian example of the auditor-exit signal, and it is directly relevant to a startup director. It is kept here, briefly, for that reason.

Timeline

• Deloitte, as auditor, forced revenue-recognition changes in the FY21 accounts (deferral of revenue on subscriptions and financed sales that had been booked up front).
• FY22 accounts were delayed by about 19 months.
• June 2023: Deloitte resigned mid-term, citing the delay in financial statements and records not being provided. Three board members quit at the same time (The Week, 22 Jun 2023).
• September 2024: the successor auditor BDO (MSKA & Associates) also resigned, flagging concerns including a transaction with a Dubai-based reseller (Business Today, 7 Sep 2024).
• Term-loan covenant breaches with its US lenders ran in parallel.

Mechanism Revenue recognised before the service is delivered, or on sales financed by third parties with recourse, inflates current revenue and receivables relative to cash: high DSRI and high TATA in Beneish terms. The auditor is the only outsider who sees the ledgers. When an auditor resigns mid-term citing missing information, the information the market would need to value the business is, by the auditor's own statement, unavailable. Since there is no reliable $F_0$, there is no valuation.