The cost of ignoring horizon in factor investing

Buying small or undervalued stocks has long been one of the most popular investment strategies. But does the risk of these strategies change depending on whether you plan to hold for one year or for five? Using 160 years of U.K. stock market data, we find that the risk profile of these strategies shifts significantly with the investment horizon, and that ignoring this can leave long-term investors measurably worse off.

This article was written by Jan Sandoval, Assistant Professor in the Department of Economics, Finance and Accounting at Esade and a member of the Group for Research in Economics and Finance (GREF).

Size and value factors

Certain types of stocks tend to outperform others over time. Small companies tend to deliver higher returns than large ones, a pattern known as the size factor. And stocks that trade at low prices relative to what the underlying business is worth tend to beat their more expensive peers, a pattern known as the value factor.

These two factors are among the most widely studied in finance. They are also at the heart of a widespread approach known as factor investing, in which investors systematically tilt their portfolios toward stocks with characteristics that have historically delivered higher returns. Today, more than two trillion dollars are managed following this approach.

Yet most of what we know about these factors comes from relatively short datasets, typically covering only the postwar period in the United States. This is a problem because measuring risk over long horizons requires long samples: with 50 years of monthly data, you have 600 observations to estimate monthly risk, but only ten for five-year risk.

In recent research with Stig Lundeby and Jens Kvaerner (both at BI Norwegian Business School), we tackle this problem using a newly constructed dataset of individual U.K. stocks spanning 1860 to 2019 and covering more than 8,000 unique stocks. The U.K. was the largest stock market in the world during much of the 19th and early 20th centuries, and the 160-year span of the data allows us to measure how risk behaves across longer horizons.

Risk changes with the investment horizon

A natural starting point when thinking about portfolio risk is that if a strategy has a certain level of risk over one month, its risk over longer periods should grow proportionally. Under this logic, the risk of holding an investment for several years would simply be a scaled-up version of its one-month risk.

This does not hold for size and value strategies. When measured directly over multi-year horizons, their risk is substantially larger than what this scaling would predict. For the value factor, long-horizon risk is roughly twice as large; the size factor shows a similar, though less pronounced, pattern.

This gap arises because returns are persistent. Periods of strong performance tend to be followed by further strength, and weak periods by continued weakness. As a result, returns are positively correlated over time, and risk accumulates faster than the scaling benchmark implies. For an investor holding a value or size strategy over several years, uncertainty is therefore greater than one-month risk estimates suggest.

The cost of ignoring horizon

How much does this matter in practice? The dataset allows us to answer this directly. We calibrate the strategies using the first half of the sample and evaluate them on the second, avoiding hindsight.

At a five-year horizon, an investor who uses horizon-appropriate risk estimates is about 11% better off than one who relies on monthly risk estimates. Equivalently, the horizon-blind investor would need 11% more starting capital to achieve the same outcome.

What does this mean for investors?

Two investors holding the same factor strategy but planning to hold it for different lengths of time are not taking the same amount of risk. A three-year investor and a ten-year investor face different risk profiles, even if the short-term statistics look identical. Portfolio allocations should reflect this. When factor returns persist over time, the optimal mix shifts with the horizon, with more aggressive tilts toward size and value at shorter horizons and a gradual move toward broader market exposure at longer ones.

A broader lesson

Time horizon is not just a preference; it is a fundamental input to how portfolios should be built. The way we measure risk matters just as much as the strategies we choose to invest in.