RiskPerf.jl

Quantitative risk and performance analysis package for financial time series powered by the Julia language.

RiskPerf.jl is engineered for high performance: core metrics avoid temporary allocations, exploit SIMD-friendly loops, and specialize on scalar vs. vector inputs. Compared with naïve broadcast-based implementations, typical routines (e.g. Sharpe/Information ratios, partial moments, summary statistics, etc.) are 5–50× faster and allocate virtually nothing, accelerating large backtests and real-time analytics.

Documentation: https://rbeeli.github.io/RiskPerf.jl/stable/

Functions

simple_returns(prices::AbstractVector; drop_first=false, first_value=NaN)
simple_returns(prices::AbstractMatrix; drop_first=false, first_value=NaN)

log_returns(prices::AbstractVector; drop_first=false, first_value=NaN)
log_returns(prices::AbstractMatrix; drop_first=false, first_value=NaN)

total_return(returns::AbstractVector; method=:simple)
total_return(returns::AbstractVector; method=:log)

cagr(returns::AbstractVector, periods_per_year::Real; method=:simple)

annualized_return(returns::AbstractVector, periods_per_year::Real)

volatility(returns::AbstractVector; multiplier=1.0)

information_ratio(asset_returns::AbstractVector, benchmark_returns::Real; multiplier=1.0)
information_ratio(asset_returns::AbstractVector, benchmark_returns::AbstractVector; multiplier=1.0)

jensen_alpha(asset_returns::AbstractVector, benchmark_returns::Real; risk_free=0.0)
jensen_alpha(asset_returns::AbstractVector, benchmark_returns::AbstractVector; risk_free=0.0)

modified_jensen(asset_returns::AbstractVector, benchmark_returns::Real; risk_free=0.0)
modified_jensen(asset_returns::AbstractVector, benchmark_returns::AbstractVector; risk_free=0.0)

drawdowns_pct(returns::AbstractVector; compound=true)

drawdowns_pnl(pnl::AbstractVector)

max_drawdown_pct(returns::AbstractVector; compound=true)

max_drawdown_pnl(pnl::AbstractVector)

capm(asset_returns::AbstractVector, benchmark_returns::Real; risk_free=0.0)
capm(asset_returns::AbstractVector, benchmark_returns::AbstractVector; risk_free=0.0)

sharpe_ratio(returns::AbstractVector; multiplier=1.0, risk_free=0.0)

adjusted_sharpe_ratio(returns::AbstractVector; multiplier=1.0, risk_free=0.0)

sortino_ratio(returns::AbstractVector; multiplier=1.0, MAR=0.0)

tracking_error(asset_returns::AbstractVector, benchmark_returns::Real; multiplier=1.0)
tracking_error(asset_returns::AbstractVector, benchmark_returns::AbstractVector; multiplier=1.0)

treynor_ratio(asset_returns::AbstractVector, benchmark_returns::Real; multiplier=1.0, risk_free=0.0)
treynor_ratio(asset_returns::AbstractVector, benchmark_returns::AbstractVector; multiplier=1.0, risk_free=0.0)

calmar_ratio(returns::AbstractVector, periods_per_year::Real; compound=true)

omega_ratio(returns::AbstractVector, target_return::Real)

skewness(x::AbstractVector; method=:moment)
skewness(x::AbstractVector; method=:fisher_pearson)
skewness(x::AbstractVector; method=:sample)

kurtosis(x::AbstractVector; method=:excess)
kurtosis(x::AbstractVector; method=:moment)
kurtosis(x::AbstractVector; method=:cornish_fisher)

upside_potential_ratio(returns::AbstractVector, threshold::Real; method=:full)
upside_potential_ratio(returns::AbstractVector, threshold::AbstractVector; method=:full)
upside_potential_ratio(returns::AbstractVector, threshold::Real; method=:partial)
upside_potential_ratio(returns::AbstractVector, threshold::AbstractVector; method=:partial)

value_at_risk(returns::AbstractVector, α::Real; method=:historical, multiplier=1.0)
value_at_risk(returns::AbstractVector, α::Real; method=:gaussian, multiplier=1.0)
value_at_risk(returns::AbstractVector, α::Real; method=:cornish_fisher, multiplier=1.0)

expected_shortfall(returns::AbstractVector, α::Real; method=:historical, multiplier=1.0)
expected_shortfall(returns::AbstractVector, α::Real; method=:gaussian, multiplier=1.0)
expected_shortfall(returns::AbstractVector, α::Real; method=:cornish_fisher, multiplier=1.0)

mean_excess(x::AbstractArray, y::Real)
mean_excess(x::AbstractArray, y::AbstractArray)

std_excess(x::AbstractArray, y::Real; corrected=true)
std_excess(x::AbstractArray, y::AbstractArray; corrected=true)

lower_partial_moment(returns::AbstractVector, threshold::Real, n::Int, method=:full)
lower_partial_moment(returns::AbstractVector, threshold::AbstractVector, n::Int, method=:full)
lower_partial_moment(returns::AbstractVector, threshold::Real, n::Int, method=:partial)
lower_partial_moment(returns::AbstractVector, threshold::AbstractVector, n::Int, method=:partial)

higher_partial_moment(returns::AbstractVector, threshold::Real, n::Int, method=:full)
higher_partial_moment(returns::AbstractVector, threshold::AbstractVector, n::Int, method=:full)
higher_partial_moment(returns::AbstractVector, threshold::Real, n::Int, method=:partial)
higher_partial_moment(returns::AbstractVector, threshold::AbstractVector, n::Int, method=:partial)

downside_deviation(returns::AbstractVector, threshold::Real; method=:full)
downside_deviation(returns::AbstractVector, threshold::AbstractVector; method=:full)
downside_deviation(returns::AbstractVector, threshold::Real; method=:partial)
downside_deviation(returns::AbstractVector, threshold::AbstractVector; method=:partial)

upside_deviation(returns::AbstractVector, threshold::Real; method=:full)
upside_deviation(returns::AbstractVector, threshold::AbstractVector; method=:full)
upside_deviation(returns::AbstractVector, threshold::Real; method=:partial)
upside_deviation(returns::AbstractVector, threshold::AbstractVector; method=:partial)

relative_risk_contribution(weights::AbstractVector, covariance_matrix::AbstractMatrix)

Bug reports and feature requests

Please report any issues via the GitHub issue tracker.

Acknowledgements

This package was inspired by the R package PerformanceAnalytics of Peter Carl and Brian G. Peterson.