Decoding Success: Jeremy Raper & The Toshiba Tussle [Part 3]
From accounting fraud to nuclear disaster: Trading a corporate implosion
Introducing Jeremy Raper
Welcome back to the Decoding Success series. Today we’re diving into part 2 of our deep dive on Jeremy Raper, the legendary investor behind Raper Capital who achieved a staggering 130x return on his portfolio over 11 years. Today’s article will proceed slightly differently than my normal articles, as I will not be a prisoner of my own format.
After reading Jeremy’s thesis into shorting Toshiba, but especially after seeing all his updates after the pitch where he closed, opened and took different sides of the trades after news flow came in, I believed this pitch and set of trades deserved its own article. Also because it is a great showcase of how a once titan of industry can get into trouble.
All pitches, I will cover on the blog in the coming weeks.
📖 Covered in Part 1
Elizabeth Arden – Short – June 2014
Aeropostale – Short – July 2014
GT Advanced Technology – Short – October 2014
Quicksilver Corp – Short – December 2014
Tuesday Morning – Short – February 2015
Chegg – Short – July 2014 (Selected "mistake pitch")
📖 Covered Part 2
Avolon – Long – March 2015
Afren Plc – Short – March 2015
Aercap – Long – October 2015
Peabody Energy – Short – March 2016
Universal Entertainment – Long – May 2016
📍 Covered in Today's article
Toshiba Corp – Short – June 2016
🔮 Coming in Future Articles
Sharp Corporation – Short – July 2017
Japan Display – Short – July 2017
Rezidor – Long – January 2018
Maxwell Technologies – Short – February 2019
Nio – Short – March 2019
Hexo – Short – October 2019
K+S Aktiengesellschaft – Short – November 2019
Tupperware bonds – Short – February 2020
Endor – Long – May 2020
Metlifecare NZ – Long – May 2020
Haier equity arb – Spread – December 2020
Automated Banking Services – Long – January 2021
Harbor Diversified – Long – February 2021
Hunter Douglas – Long – April 2021
Cardno – Long – September 2021
FAR Limited – Long – February 2022
Shell Midstream – Long – June 2022
Twitter #1 – Long – July 2022
Twitter #2 – Long – September 2022
Danakali – Long – October 2022
Montero Mining – Long – February 2024
In April 2016, while most investors were running from Toshiba’s recent accounting scandal, Jeremy Raper was doing the opposite, he was building a short position. The stock had already cratered from ¥400 to ¥234 following revelations of one of Japan’s largest corporate frauds, but Raper saw something the market was missing: this wasn’t over. Not by a long shot.
Against this backdrop of fraud and value destruction, Jeremy would execute a series of trades that showcased his ability to adapt as circumstances evolved. But before we can dive into his first trade, we need to understand what Toshiba actually was, what the company had been through in recent years, and how this volatile corporate situation would create a myriad of opportunities.
The Giant That Was Toshiba
Toshiba Corporation stood as one of Japan’s most storied industrial conglomerates, with roots stretching back to 1875. By the 2010s, it had evolved into a diversified multinational with significant presence across multiple high-tech sectors, particularly renowned for its pioneering work in NAND flash memory technology, a market it helped create alongside Samsung.
By the end of 2015, Toshiba operated through six major business segments:
Revenue and EBITDA Breakdown (Fiscal Year Ended March 31, 2015)
Electronic Devices & Components (25.3% of revenue) was the crown jewel, generating ¥283.9 billion in EBITDA, by far the highest of any segment. This division housed the semiconductor business (memory, discretes, system LSIs) and storage products, producing outsized margins that made it extraordinarily profitable.
Energy & Infrastructure (28.3% of revenue) was the largest segment, encompassing nuclear power systems, thermal and hydro power, and transmission infrastructure. Despite its revenue scale, this capital-intensive division contributed far less to EBITDA due to thinner margins and slower profit recognition on long-term projects.
Community Solutions (14.3% of revenue) covered elevators, lighting, and building systems with healthy margins. Healthcare Systems (8.8% of revenue) produced medical equipment and generated solid returns. Lifestyle Products & Services (16.6% of revenue) was already struggling with heavy competition in consumer electronics, posting a concerning -¥98.2 billion EBITDA loss.
The 2015 Accounting Scandal: A Culture of Deception
In July 2015, Toshiba admitted to systematically overstating profits by over ¥220 billion (approximately $1.9 billion) across seven years. This wasn’t a rogue accountant, the manipulation spanned all major segments, revealing a culture where meeting targets mattered more than honest reporting.
The most troubling pattern emerged in Energy & Infrastructure, where long-term project costs were systematically understated. For projects spanning years, like nuclear power plant construction, management had enormous discretion over cost recognition. That discretion had been systematically abused.
CEO Hisao Tanaka, Chairman Atsutoshi Nishida, and Vice Chairman Norio Sasaki all resigned. The Tokyo Stock Exchange placed Toshiba on a supervision watchlist, threatening delisting. The stock crashed from ¥400 to below ¥200.
Toshiba Stock price chart from early 2015 to 2016
Asset Sales (March 2016)
With its balance sheet under severe stress and delisting threats looming, Toshiba moved quickly to raise capital. In March 2016, the company announced two major divestments.
Toshiba Medical Systems Corporation was sold to Canon Inc. for ¥665.5 billion. Then China’s Midea Group acquired an 80.1% stake in Toshiba Lifestyle Products & Services Corporation for ¥53.7 billion, taking over the struggling home appliances business along with 40-year global rights to the Toshiba brand.
These sales brought in over ¥700 billion in much-needed capital. But would this be enough to stabilize the company?
April 20 Trade 1 (Short): The Initial Short
Entry: April 20, 2016 at ¥234 per share
Jeremy saw opportunity in shorting Toshiba even after the stock was down big due to the accounting scandal, as recent short covering had driven the stock up again. Raper’s short thesis was built on fundamental analysis that looked beyond the immediate accounting scandal to the underlying business reality.
Jeremy’s thesis had three pillars:
1. The Balance Sheet Reality
Even after asset disposals, Toshiba’s balance sheet remained “woefully over-leveraged.” The equity ratio stood at a dangerously low 8.8%, with an alarming 235% intangibles-to-equity ratio. Raper, who had long worked in Japanese debt markets, had identified that Japanese banks were usually unwilling to refinance debt if the equity ratio fell below 10%. This meant that Toshiba was likely facing forced dilution in the near term.
2. Unrealistic Management Guidance
Raper saw through management’s optimistic forward guidance, describing the business outlook as “dire.” His sum-of-parts analysis suggested the stock was 65% overvalued, with a fair value around ¥80 per share—far below the market price of ¥234.
3. The Nuclear Time Bomb
Westinghouse, Toshiba’s US nuclear subsidiary, had problems far from resolved, creating a looming liability that could devastate the company. Westinghouse had underperformed for years and had not won a single new tender since the Fukushima disaster in 2011. Meanwhile, Toshiba had still never impaired the large goodwill it carried for this business. Further impairing this goodwill would bring Toshiba’s equity ratio so low that no Japanese bank would be willing to lend to them anymore.
May 2016: The Full Year Report Confirms the Worst
When Toshiba published its full year 2015 results in May 2016, the numbers validated Raper’s concerns. The equity ratio came in at just 5.8%—even worse than his initial 8.8% estimate and far below the critical 10% threshold Japanese banks required for refinancing.
Management’s guidance for fiscal 2016 bordered on fantasy: across multiple segments, they projected double-digit revenue declines paired with massive margin expansion that didn’t withstand scrutiny. For Raper, the report confirmed management remained either delusional or deceptive.
September 2016: NAND Recovery Drives Stock Higher
By September, the stock had surged as the NAND/HDD memory business dramatically outperformed expectations. A massive restocking cycle driven by the smash hit iPhone 7 created huge demand for memory chips. The surge in profits from this valuable division improved Toshiba’s equity ratio to 7%, still dangerously low, but moving in the right direction.
For shorts like Raper, this was painful. The very asset he knew was valuable was now masking the problems elsewhere in the company, driving the stock higher even as fundamental issues remained unresolved.
December 28, 2016: The Nuclear Bomb Detonates
Then came the news Raper had been anticipating. Toshiba announced it would take several billion dollars in impairments on Westinghouse, driven by massive cost overruns on two existing projects. The stock fell 30% in a single day.
Based on these write-downs, Raper calculated the company would need to dilute shareholders by roughly 50% of the market cap just to shore up the balance sheet and keep Japanese banks willing to lend. But there was a problem: being on the TSE watchlist meant Toshiba likely couldn’t raise capital through a public offering.
February 14, 2017: The Memory Sale Announcement
On February 14, Toshiba delayed its earnings report, and the Nikkei reported the company would include a going concern warning in its accounts. The situation looked dire.
But then came the critical development: Toshiba publicly announced it was considering selling part or all of the memory business. Given how important and valuable this business was, the eventual sale price could be enormous.
Trade 1 Covered
Exit: February 21, 2017 at ~¥183 per share Result: 23% profit
Toshiba stock chart from around April 2016 (Initial short) till February 2017 (Covered)
As we can see from the chart, this was not an easy short to hold, with the stock nearly doubling over the period. As Raper admits himself, he misread the technicals, the short covering was not over and this led to upward pressure on the stock. However, in the end the short did work, delivering a 23% profit.
JR closed this trade in February 2017, mainly because he was very wary of an irrational buyer stepping in and buying the NAND business for too high a price tag, creating significant upside risk that would overwhelm his short thesis.
The Westinghouse Bankruptcy: Validating the Thesis
Just one month after Raper covered his initial short, his caution about the nuclear business proved devastatingly accurate. On March 29, 2017, Westinghouse filed for Chapter 11 bankruptcy, citing cost overruns that would create losses for Toshiba exceeding $9 billion, nearly three times previous estimates.
March 17 Trade 2: The Quick Re-Entry
Entry: March 17, 2017 at ¥200 per share
With Westinghouse’s bankruptcy now public and the stock falling, Raper re-entered his short position. The rationale seemed compelling, JR had the following thesis:
Complete NAND Sale Expected: Raper believed a 100% sale of the memory business was now likely, which would be “bad for long-term equity value” as it represented the company’s most valuable asset.
Additional Nuclear Losses: The Westinghouse bankruptcy created additional losses of ¥300-500 billion, further devastating the balance sheet.
Poor Remaining Business: After stripping out the memory business, what remained looked like “Hitachi, just lower quality and with huge nuclear tail risk.” His price target remained ¥120-130 per share.
April 7, 2017: FoxxConn Bid
However, the same issue that forced his first exit emerged even more dramatically. On April 7, reports surfaced that Hon Hai Precision Industry (part of Foxconn) had bid ¥3 trillion ($27 billion) for Toshiba’s memory business.
Trade 2 Covered
Exit: April 7, 2017 at ¥215 per share Result: 10% loss
Toshiba stock chart from around March 2017 (Second short) till April 2017 (Covered)
In JR’s eyes, this was a completely irrational bid at “~30x P/E and 10x EBITDA” at a high point in the always cyclical NAND industry. The valuation made no sense, but that was precisely the problem for his short position.
After the Hon Hai bid emerged, Raper quickly covered the same day. When potential upside from an irrational acquirer exceeds your downside case, the risk-reward of a short position collapses.
This trade demonstrated a crucial principle: being right about the fundamentals isn’t enough if the setup creates unlimited upside risk from unpredictable actors.
May 2017: Western Digital Throws a Wrench
Just as Toshiba was running its sales process, a major legal obstacle emerged. In May 2017, Western Digital(WD) filed for arbitration, arguing that Toshiba could not sell the NAND unit without WD’s consent under their joint venture terms.
The dispute centered on their long-standing JV operating NAND flash fabs in Yokkaichi, Japan. Western Digital claimed its consent was required for any sale affecting the JV, while Toshiba rejected this interpretation. WD pursued international arbitration to block the sale, while Toshiba countersued for obstruction.
This created enormous uncertainty: even if Toshiba found a buyer willing to pay enough to restore positive equity, Western Digital’s lawsuit could prevent the deal from closing.
June 2017: The Delisting Threat Looms
Under Tokyo Stock Exchange regulations, a company recording negative shareholders’ equity for two consecutive fiscal years faces mandatory delisting. With Toshiba’s equity having fallen to approximately -¥581 billion as of March 31, 2017 due to the Westinghouse writedowns, the company was now in serious jeopardy. If they couldn’t restore positive equity by March 31, 2019, they would be removed from the exchange entirely.
The situation was becoming a perfect storm: Toshiba needed to complete the memory business sale to generate enough proceeds to solve the equity problem and avoid delisting. But Western Digital’s lawsuit was preventing the deal from closing.
June 20, 2017: Bain Consortium Selected
On June 20, 2017, Toshiba chose a consortium led by Bain Capital, backed by INCJ, Development Bank of Japan, Apple, and other Japanese entities, as its preferred bidder, valuing the memory unit at approximately ¥2.1 trillion (roughly $19 billion).
Notably, this bid was lower than Foxconn’s ¥3 trillion offer, but Toshiba preferred the Bain consortium because it included Japanese government entities rather than Chinese buyers, reducing regulatory concerns. If completed, the deal would generate enough proceeds to restore Toshiba’s equity to positive territory, solving the delisting crisis.
But there was a critical problem: Western Digital’s lawsuit was still blocking the transaction from closing.
Trade 3: Betting Against Delisting Risk
Entry: July 24, 2017 at ~¥300 per share
By summer 2017, with the stock having recovered to ¥300, Raper identified delisting risk as a new catalyst for shorting. The disputes between Western Digital and Toshiba over the memory business sale created genuine risk that Toshiba could face “snap delisting.”
Even though Toshiba had selected the Bain consortium as the preferred bidder, WD’s arbitration was blocking the transaction. Without the sale closing, Toshiba couldn’t restore positive equity, and the delisting clock kept ticking toward the March 2019 deadline. Raper cited the Seibu Rail precedent, where delisting had caused the stock to fall dramatically. He expected Toshiba to fall to ¥150 if delisting was announced.
August 1, 2017: TSE Makes It Official
On August 1, 2017, the Tokyo Stock Exchange officially downgraded Toshiba from the prestigious First Section to the Second Section and placed the company in a “grace period pertaining to delisting.” The deadline was set: Toshiba had until April 1, 2019 to restore positive shareholders’ equity or face removal from the exchange.
November 2017: The Private Placement Solution
In November 2017, Toshiba’s adviser Goldman Sachs pulled off a remarkable feat: selling ¥600 billion (about $5.3 billion) of new shares through a third-party allotment to foreign investors. This private placement circumvented a critical problem—due to ongoing audit concerns, Toshiba could not simply offer new shares on the public market.
The placement plugged the negative equity hole, ensuring Toshiba would not be forced to delist. Remarkably, the shares were priced only 10% below prevailing market prices, minimizing the immediate stock drawdown despite the new shares representing about 54% of the existing share count, one of the largest placements and dilutions in Japanese corporate history.
Trade 3 Covered
Exit: November 20, 2017 at ¥275-280 per share Result: 6% gain
Toshiba Stock Chart from around June 2017(Third short) till November 2017(Covered)
Toshiba’s November 2017 equity offering “clearly removes delisting risk,” according to Raper. The massive private placement had successfully restored positive equity, satisfying both regulators and Japanese banks.
While he still believed “the core business is still terrible and wildly overvalued,” the offering created “very significant risk that Toshiba ends up owning either all the memory business or a much bigger chunk of it”, fundamentally changing the risk-reward profile. With the Western Digital lawsuit still unresolved, there was now a real possibility the memory sale might not complete, meaning Toshiba would retain its most valuable asset.
Once again, Raper demonstrated the discipline to exit when his specific catalyst was invalidated, even if he remained negative on the underlying business.
December 13, 2017: Peace with Western Digital
On December 13, 2017, Toshiba and Western Digital announced they had dropped all lawsuits and arbitration proceedings, agreeing to renew their joint venture investment deals through 2027. The legal standoff that had blocked the memory business sale for months was finally resolved.
This development opened the door for the Bain Capital-led consortium deal to move toward final closing, removing the last major obstacle to the ¥2 trillion transaction.
Trade 4: Finally Long
Entry: April 6, 2018 at ~¥299 per share
It’s remarkable that after shorting Toshiba three times, Raper managed to convince himself to go long. But by early 2018, he had identified a compelling opportunity based on the economics of the memory business sale.
The original Bain Capital consortium deal had been struck in June 2017, valuing the memory business at ¥2.1 trillion when memory earnings were deeply depressed. The deal was supposed to close in March 2018, but that timeline had lapsed without completion, due to the Chinese regulator not approving the deal yet. Crucially, this meant Toshiba was now free to renegotiate or recut the deal entirely.
By early 2018, the memory business had transformed. The unit’s EBIT had surged from roughly ¥150 billion when the deal was negotiated to approximately ¥450 billion, a massive improvement driven by the cyclical recovery in memory pricing. Raper calculated that the Bain consortium was getting a “$10-15 billion windfall” simply due to timing.
The economics were stark: any rational board would renegotiate when the underlying asset had appreciated so dramatically. Extracting even a portion of this windfall could add ¥1 trillion or more in value, representing over 50% upside on Toshiba’s ¥1.8 trillion market cap. The risk-reward looked compelling: limited downside if the deal proceeded as agreed, but massive upside if management acted rationally.
Adding to the thesis, foreign activist investors were taking significant stakes and pushing for better governance and value extraction. The pressure for management to renegotiate was building.
May 15, 2018: Strong Earnings Reinforce the Thesis
When Toshiba reported earnings on May 15, 2018, the results validated Raper’s analysis. The company generated ¥530 billion in EBIT, of which ¥465 billion came from the memory business. At current prices, Toshiba traded at roughly 5x EV/EBIT—a reasonable valuation for what remained.
More importantly, the balance sheet had significantly strengthened. There was no longer any financial imperative to sell the memory business at all. Toshiba could afford to keep it or negotiate a much better price.
The company did announce they still intended to close the memory sale, and if completed, they would execute a major share buyback, potentially repurchasing 30% of outstanding shares. But the lack of urgency created an opening for renegotiation.
June 1, 2018: The Deal Closes Without Renegotiation
On June 1, 2018, the deal finally closed after China, the last major jurisdiction, granted merger approval. The transaction was effectuated at the original ¥2 trillion price, with no renegotiation despite the dramatic change in the memory business’s profitability.
Raper described this as “one of the most egregious breaches of fiduciary duty the capital markets may have ever seen.” The traditional Japanese corporate culture of honoring agreements, even when circumstances had changed so dramatically that renegotiation would have been economically rational, prevailed over shareholder value maximization. Management left ¥1 trillion or more on the table.
The stock performed well nonetheless. By July 2018, it was up roughly 20% while the broader market was flat and many tech names had been destroyed. The drivers were under-ownership and new management’s commitment to buy back ¥700 billion of stock in the second half of the year, about 30% of the market cap.
Trade 4 Reduced/Closed
Exit: July 21, 2018 at ~¥350 per share (reduced position significantly) Result: ~17% gain
Toshiba stock chart from around April 2018 (Long) till July 2018 (Position reduced)
On July 21, 2018, at ¥350, Raper reassessed the risk-reward. With the windfall memory renegotiation now off the table, the setup had fundamentally changed. He thought the stock could still test ¥400 due to technical buying pressure from the buyback and ongoing low expectations. But new risks had emerged:
The ¥700 billion buyback could be executed through a tender offer off-exchange rather than gradual market purchases, removing the technical support
The new mid-term plan management was developing could still disappoint
NAND prices were clearly rolling over, threatening future profitability
It remained unclear exactly how much of the memory business Toshiba would retain going forward
The valuation support, in the absence of the memory deal renegotiation optionality, simply wasn’t there anymore. Raper concluded the stock was still “somewhat under-owned/misunderstood,” justifying keeping a small position. But he took most of his chips off the table, significantly reducing his long position.
Final Position and Reflection
By November 2018, Raper maintained a reduced long position, benefiting from the buyback program and activist involvement. His final updates showed him still holding shares but with a much more cautious stance given the improved valuation and reduced upside optionality.
The Complete Trading Record
Over roughly two years, Raper executed four distinct trades on Toshiba, demonstrating remarkable adaptability:
Trade 1 (Short) - Weak balance sheet and looming nuclear losses → +23% profit
Trade 2 (Short) - NAND sale likely bad for long-term equity value, additional nuclear losses → -10% loss
Trade 4 (Long) - Optionality of memory deal being renegotiated → +17% profit
Stringing these trades together delivered a cumulative return of approximately 37% over just a bit more than two years, a strong performance navigating one of Japan’s most turbulent corporate crises. More importantly, Raper demonstrated the intellectual flexibility to switch sides multiple times as circumstances evolved, not letting prior positions dictate future decisions.
If you made it al the way to the end please subscribe, for more future content
![Decoding Success: Jeremy Raper [Part 2]](https://substackcdn.com/image/fetch/$s_!ujPu!,w_140,h_140,c_fill,f_webp,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9e1a94bf-d242-4073-bfb5-772352247a9a_1920x1080.png)
I really enjoy this series so far about JR. Thanks for writing.
Nice article!