Building CetaResearch. Data analysis for policy, research, and startups.
DCF
We ran the DCF Threshold screen on 25 years of Indian stocks from BSE and NSE. The signal: buy the 30 most undervalued companies by Gordon Growth DCF discount, filtered for ROE above 8%, D/E below 1.5, and positive operating cash flow. The strategy found no qualifying stocks
DCF
We ran the DCF Threshold screen on 25 years of Canadian stocks from the TSX. The signal: buy the 30 most undervalued companies by Gordon Growth DCF discount, filtered for ROE above 8%, D/E below 1.5, and positive operating cash flow. The strategy returned 9.32% annually vs
event study
Pre-earnings drift across 14 markets with local currency benchmarks. Korea leads at +0.709% vs KOSPI, India +0.587% vs Sensex. Japan's effect disappeared vs Nikkei. Taiwan flipped from negative to positive. Benchmark choice matters.
event study
15,216 JPX earnings events from 2017-2025. When benchmarked against the Nikkei 225 instead of EWJ, the pre-earnings drift drops to -0.014% (t=-0.36). Not significant. The original +0.114% result was a benchmark artifact.
event study
163,661 US earnings events from 2000-2025. Habitual beaters drift +0.266% vs S&P 500 in 10 days before announcements. Habitual missers fall -0.479%. The spread between best and worst beat-rate quintiles is 53bp.
event study
6,626 KOSPI earnings events from 2016-2025. Pre-earnings drift averages +0.709% vs KOSPI in 10 days before announcements. Both habitual missers (+0.610%) and mixed stocks (+0.728%) show the same drift. Beat rate doesn't matter in Korea.
sector rotation
Same sector mean reversion strategy across 14 exchanges. India +8.40%, Korea +6.63%, Taiwan +6.58% excess CAGR. Germany -3.20%, Switzerland -3.51%. 12-point spread between best and worst. The split is geographic and explainable.
backtests
Sweden is one of two dedicated markets in our study with positive excess returns. 9.67% CAGR (+0.78% excess), Sharpe 0.344, high variance. The Swedish industrial economy fires the expansion signal less often but delivers strong results when invested.
backtests
Germany produced the best risk-adjusted cyclical timing result among developed markets. 8.48% CAGR (-0.41% excess), Sharpe 0.411, max drawdown -35.07%. German industrials and materials provided stable expansion signals across 24 years.
EV/EBITDA
We tested a sector-relative EV/EBITDA screen on 25 years of US stock data. When a profitable company's EV/EBITDA falls 30%+ below its sector peers, it tends to mean-revert. 12.34% CAGR vs 7.64% for SPY, with 62.9% down capture and zero cash periods across 25 years.
UK stocks
We backtested a sector-relative EV/EBITDA screen on 25 years of UK stock data (LSE). 10.35% CAGR vs 7.64% for SPY, zero cash periods, and a max drawdown of -31.12%. The signal worked every year on the LSE — consistent, not spectacular.
Switzerland stocks
We backtested sector-relative EV/EBITDA on 23 years of Swiss data (SIX). 9.32% CAGR vs 7.64% SPY, but the real story is the downside protection: 35.3% down capture, MaxDD -29.61%, and a Sharpe ratio of 0.467, matching the US on risk-adjusted terms with a 15-stock portfolio.