A Korean media company lost ₩783bn before tax over three years and still booked ₩178bn of tax expense

I read DART filings every day. This one made me stop.

CJ ENM is the company behind Parasite and a lot of Korean television. It shows up on screens at 998 times earnings. My first assumption was a data error. It isn't. Trailing twelve-month net income is ₩738 million. Million, on ₩5.2 trillion of revenue. Divide a ₩0.7 trillion market value by a rounding error and you get 998.

The operating line is fine. It's everything below it that eats the company.

Take 2025, the year that looks healthy. Operating income ₩132.9bn on revenue of ₩5,134.5bn. Pretax income ₩84.2bn after financing costs. Income tax expense ₩67.2bn. Net income ₩17.0bn. That's an effective rate near 80% in a profitable year.

The two years before it are stranger. In 2024 the group lost ₩531.9bn before tax and recorded ₩48.9bn of tax expense. In 2023 it lost ₩335.1bn before tax and recorded ₩61.7bn. Add the three years: ₩782.9bn of cumulative pretax losses, and ₩177.8bn of tax expense on top of them.

Two things are going on. The annual figures don't separate them cleanly.

The first is structural. It's the part I think US readers miss. A consolidated income statement adds up profits and losses across a group. Tax authorities generally don't. A profitable subsidiary pays tax on its own profit even if an affiliate lost money that year. So a group can report a consolidated loss and a real tax bill in the same year. If you're used to reading a US consolidated return, seeing those two coexist this violently is odd.

The second is deferred tax. In 2022 the tax line was a ₩150.8bn benefit. That is what recognising deferred tax assets looks like. In the third quarter of 2024 the company booked ₩130.5bn of tax expense. It had a ₩400.9bn pretax loss that quarter. That is what derecognising them looks like. You write those assets off when you stop expecting enough future profit to use the accumulated losses. If that's what happened, management said something about its own forecasts that no press release did.

The cash is real, by the way. Operating cash flow was ₩1,166.4bn in 2025. It was ₩1,402.6bn in 2024 and ₩1,294.8bn in 2023. That's ₩3,863.8bn over three years against ₩960.6bn of cumulative net losses. Investing absorbed ₩1,023.2bn in 2025, about 88% of the operating inflow. Only ₩29.4bn of that was property and equipment.

That's the signature of a content business. The business capitalises programming when it makes it. It amortises programming when it airs. So amortisation lifts operating cash flow while the production spend sits in investing. The cash arrives. The business recycles it into next year's slate. Accounting profit lands somewhere in between.

Which makes the two headline multiples less exciting than they look. The 998x is arithmetic on a nearly-zero denominator. The 0.21 price-to-book isn't the market calling the equity worthless either. Intangibles are ₩2,302.5bn of ₩3,513.6bn of equity. Stripping them out puts tangible book near ₩1,211.1bn, or roughly 0.6 times.

Figures are from the DART filings; the Korean originals govern.

No position.

I can't get the yearly split between cash tax actually paid and deferred movement from the annual reports. The company discloses the total. It doesn't disclose the split. For a company like this the split is the whole question. If you read Korean filings regularly and know where that breakdown lives, tell me.

(Written with AI help. I pulled the figures from the filings myself and checked them; the drafting used an AI model.)