This post summarizes part of Quant Research Team Report from QRAFT Technologies. For more information, please subscribe to newsletter published by QRAFT Technologies
본 포스팅은 QRAFT Technologies의 Quant Research Team Report의 일부를 요약한 것입니다. 이에 대해 더 자세히 알고 싶다면 QRAFT Technologies에서 발간하는 뉴스레터를 구독해주세요.
(Subscribe Link : https://www.qraftec.com/newsletter)

Introduction

From 1950s, Intangible capital propotion in companies have been growing. The reason is that the industry structure is changing from secondary industry such as manufacturing, where tangible assets are important, to more advanced industry, where intangible assets such as IT technologies, patents creates more value. [Figure 1] below shows the trend of the intangible assets proportion in the recent total assets. We can observe that the overall weights have become much larger than the past.

[Figure 1. Intangible asset weights trend]
* Mean value for each period calculated from stocks of companies listed in NYSE, NASDAQ, AMEX.

Noteworthy part is the cost of the R&D, where new products and technological advances result from. R&D costs are the costs that companies spend on activities to develop new products or technologies and are most often recorded as expenses in accounting standards. However, since R&D costs are not directly linked to profitability, it is necessary to check using RORC indicators whether actual R&D costs and revenue are linked.

Equation of RORC

According to the above equation, when the R&D cost increases, RORC can increase only if the gross profit for the year follows. [Figure 2] on the right shows that the RORC trend has slightly increased compared to the past. This shows that the R&D costs of companies have been better converted to profits more recently.

[Figure 2. Trend of average R&D expenses and RORC]
* Mean value for each period calculated from stocks of companies listed in NYSE, NASDAQ, AMEX.

Most companies expect to create corporate value while building external and internal moat by spending R&D expenses. However, R&D literally becomes a wasted cost if companies cannot turn it into actual profit. Unnecessary expenses negatively affect the company’s financial structure and value, and high level of R&D expenses can be considered risks. Therefore, the R&D costs can be either the appropriate investment or a risky adventure depending on the company’s external structure even if the amounts are the same. In this post, we construct a portfolio based on R&D Capital to Asset (Rca) rather than pure R&D cost from companies, measured by dividing R&D cost by total assets. We then perform empirical analysis on whether Rca is a factor that leads to a real meaningful return.

Anomaly Historical Performance

The sample period for measuring performance is from January 1992 to April 2020, and the separate analysis for equal-weighted and market-cap-weighted portfolios are presented.

[Table 1. Portfolio Performance]

[Table 1] above shows the performance of the quintile portfolio based on the Rca factor. In the equally weighted portfolio, we can observe that the Ann CAGR and Ann Sharpe Ratio increase as we see higher quantile, and we can also observe that the market cap weighted portfolio has a similar pattern.

[Table 2. IC Table]

*** p-value < 0.01, ** p-value < 0.05, * p-value < 0.10
Source : QRAFT Technologies, Compustat

[Table 3. Alpha Result]

Next, let us look at the significance of IC measurements and Alpha to measure factor robustness. The IC is calculated using Pearson or Spearman correlation between the factor value at time t-1 and the stock return at time t. If the IC value derived from continuous values, it is called just the IC, while ranked comparison is called the rank IC. The IC and Rank IC values in [Table 2] were all significant, and we may argue that the Rca factor has predictive power over the next period return. In addition, [Table 3] shows that all the alpha values were significant except CAPM alpha, when the excess return, CAPM alpha, and Fama-French 3 Factor alpha were tested using the Newey-West t-test with a delay time difference of 12.

Recent Performance & Strategy with Anomaly

[Figure 3. Portfolio Performance for Half, Tercile, Quintile, Decile(left), Portfolio Performance Recent 3, 5, 10 year(right)]

[Figure 5(left)] shows the results of portfolio returns based on the Rca factor for various quantiles and long-short portfolio returns from January 2010 to April 2010. We observe that the level of return is rising, therefore, it may be enough to conclude that the relationship between factor scores and returns is monotonic.

Also, in [Figure 5(Right)], within the same logic, more recent data shows clearer return deviation between the highest and the lowest quantile. Therefore, we can say that the significance of Rca factor has become larger.

[Table 4. Recent Period IC Table]

[Table 4] shows IC tables for the last 3, 5, and 10 years. The coeffficents for each IC values are significant throughout the whole period, and the predictive power has become larger as we observe more recent period. The results confirm that both the Long-Only, Long-Short strategy using Rca factor can work robustly.

Strategy with Rca Anomaly

As the robustness of the factor has remained strong recently, long-only and long-short strategies may be implemented using Rca. For long-only strategy, we invest in the group ranked in the top 10% based on the Rca factor using either equal weights, or market capital weights.

For long-short strategy, we long top 10% stocks and short bottom 10% stocks based on Rca factor. At the same time, we propose a strategy to long top quantile stocks and short the benchmark index to aim only market excess-return generated by the factor. [Figure 4] and [Figure 5] below show the performance of each strategy over the last 10 years.

[Figure 4. Equal-weighted strategy back test result]

[Figure 5. Market cap weighted strategy back test result]

R&D Capital To Asset(Rca Anomaly)

Introduction

Anomaly Historical Performance

Recent Performance & Strategy with Anomaly

Strategy with Rca Anomaly

Related Articles

Search for More Articles