Total-Value-Locked and Active Developers
Table of Contents
Using TVL sorted portfolios, Brigida (2025)1 found no evidence that TVL can generate alpha. Returns on TVL-sorted portfolios are linear functions of crypto market returns. These results also imply TVL-sorted portfolios do not proxy an investment intensity or value effect. Therefore, in a series of blog posts I'll test whether portfolios sorted on both TVL and other factors can generate alpha. Specifically, I test the multisort portfolios for significant alpha in regression on excess crypto market returns, and a crypto size factor.
There are a number of reasons to think a cryptocurrency's TVL will affect its returns. First, because total-value-locked (TVL) measures investment in a cryptocurrency's ecosystem, it is often viewed as a measure of trust. Greater trust, it is thought, should increase the cryptocurrency's price.
In addition, TVL may also affect prices through traditional investment in value channels. Fama and French (2015)2 found firms which invest more aggressively tend to earn lower returns. Since TVL to market cap may be viewed as an investment intensity factor, there may be a similar effect in crypto. TVL to market cap may also be a value measure which signals future growth. Liu, Tsyvinski, and Wu (2022)3 found evidence of a value effect in crypto using a price to new address ratio.
1. TVL and Active Developers Multisort Portfolios
In my tests I follow the standard Fama–French 2×3 portfolio construction, adapted to crypto. Each period I rank all eligible tokens on two characteristics measured in the prior period: total value locked (TVL) and the number of active developers (scaled by market cap). First, I split tokens into Small TVL and Big TVL groups using the median TVL. Bitcoin is excluded from the sample, though this does not affect the results. Independently, I sort tokens into three developer-activity groups — Low, Medium, and High devs — using the 30th and 70th percentiles of the developer measure. Taking the intersection of these two independent sorts gives six portfolios: Small–Low (SL), Small–Medium (SM), Small–High (SH), Big–Low (BL), Big–Medium (BM), and Big–High (BH). I then track the value-weighted returns of these six portfolios over the next period. This 2×3 design lets me see how returns vary jointly with TVL and developer activity, and then build long–short factors (e.g., high-minus-low devs controlling for TVL) to test whether these characteristics earn a risk premium.
1.1. TVL (size) factor: Small minus Big (SMB)
The long-shot SMB portfolio measures whether TVL is a rewarded risk controlling for developers to market cap.
\[\text{SMB}_t = \frac{1}{3}\bigl(R_{SL,t} + R_{SM,t} + R_{SH,t}\bigr) - \frac{1}{3}\bigl(R_{BL,t} + R_{BM,t} + R_{BH,t}\bigr)\]
Descriptive statistics. Data is weekly, and in percentage points.
0
count 187.000000
mean 0.106147
std 5.438021
min -18.959746
25% -2.332798
50% -0.024520
75% 2.373632
max 25.173556
The portfolio as a weekly mean return of 0.11 %, however given the T-test below, the mean return is insignificantly different from 0.
TtestResult(statistic=array([0.26692505]), pvalue=array([0.78982258]), df=array([186]))
1.2. Devs factor: High devs minus Low devs (HML), neutral to TVL
This long-short portfolio is long tokens with more developers (per market cap) and short tokens with few developers, while controlling for TVL.
\[\text{DEV}_t = \frac{1}{2}\bigl(R_{SH,t} + R_{BH,t}\bigr) - \frac{1}{2}\bigl(R_{SL,t} + R_{BL,t}\bigr)\]
Descriptive statistics. Data is weekly, and in percentage points.
0
count 187.000000
mean 0.420205
std 7.548905
min -22.916503
25% -4.430381
50% -0.080681
75% 3.783944
max 27.429632
The portfolio as a weekly mean return of 0.42 %, however given the T-test below, the mean return is insignificantly different from 0.
TtestResult(statistic=array([0.7611982]), pvalue=array([0.44750268]), df=array([186]))
The TVL and Developer SMB portfolio has a mean annualized return of 9.93% and a standard deviation of 35.04%. The mean return is insignificantly different from 0. The HML portfolio has a mean annualized return of 15.70% and a standard deviation 60.42%, though again the mean return is insignificantly different from 0.
2. Joint Tests for Significant Alpha
To test whether my TVL–developer portfolios earn any abnormal returns (alpha), I use a standard linear factor model and Hansen’s J-statistic4 for a joint alpha test.
I start from a 2×3 sort on lagged total value locked (TVL) and active developers (both scaled by market cap), which gives me two test portfolios. For each portfolio, I ask whether its average excess return can be explained by two simple crypto risk factors:
- \(R_{M_t} - R_{F_t}\): excess return on the broad crypto market
- \(Market\ Cap\ SMB_t\): a "small minus big" size factor that goes long small-cap tokens and short large-cap tokens
That is, I estimate both:
\(\text{SMB}_t = \alpha_1 + \beta_{M,1}\bigl(R_{M_t} - R_{F_t}\bigr) + \beta_{S,1} Market\ Cap\ SMB_t + \epsilon_t\)
and
\(\text{HML}_t = \alpha_2 + \beta_{M,2}\bigl(R_{M_t} - R_{F_t}\bigr) + \beta_{S,2} Market\ Cap\ SMB_t + \epsilon_t\)
and then test whether \(\alpha_1\) and \(\alpha_2\) are jointly equal to 0 via Hansen's J-Statistic.
2.1. Results
TradedFactorModel Estimation Summary
================================================================================
No. Test Portfolios: 2 R-squared: 0.1478
No. Factors: 2 J-statistic: 1.4513
No. Observations: 187 P-value 0.4840
Date: Sat, Dec 27 2025 Distribution: chi2(2)
Time: 10:56:15
Cov. Estimator: kernel
Risk Premia Estimates
========================================================================================
Parameter Std. Err. T-stat P-value Lower CI Upper CI
----------------------------------------------------------------------------------------
excess_crypto_market -0.0686 0.0063 -10.882 0.0000 -0.0810 -0.0563
single_smb_factor -1.2948 1.0534 -1.2292 0.2190 -3.3594 0.7698
========================================================================================
Covariance estimator:
KernelCovariance, Kernel: bartlett, Bandwidth: 4.0
See full_summary for complete results
See the results above, estimated via the Traded Factor Model method from the LinearModels python package. The J-Statistic (1.45) is a test of whether the alpha coefficients are jointly equal to 0 (the null hypothesis). We can't reject the null with the statistics p-value of 0.48. We therefore conclude that portfolios sorted both on TVL to market cap and the scaled number of developers does not generate alpha. That is, these portfolio returns can be replicated by appropriate weights on the crypto market and SMB factor portfolios.
2.1.1. Annualized Alpha Coefficients
The SMB and HML portfolios generate an alpha of 35.36 %, and 45.76 %, respectively. However, given large standard errors, these alphas are insignificantly different from 0.
multi_SMB 35.131173 multi_HML 45.763545 Name: alpha, dtype: float64
Footnotes:
Brigida, Matthew (2025), "The surprising irrelevance of total-value-locked on cryptocurrency returns," Economics Letters, Vol 257.
Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of financial economics, 116(1), 1-22.
Liu, Y., Tsyvinski, A., & Wu, X. (2022). Common risk factors in cryptocurrency. The Journal of Finance, 77(2), 1133-1177.
Hansen, Lars Peter (1982), "Large Sample Properties of Generalized Method of Moments Estimators," Econometrica.