{"id":27874,"date":"2013-04-04T02:01:00","date_gmt":"2013-04-04T00:01:00","guid":{"rendered":"http:\/\/beta.next-finance.net\/strategie\/risk-factors-taking-risk-budgeting-one-step-further\/"},"modified":"2019-12-30T22:34:00","modified_gmt":"2019-12-30T21:34:00","slug":"risk-factors-taking-risk-budgeting-one-step-further","status":"publish","type":"post","link":"http:\/\/beta.next-finance.net\/en\/strategie\/risk-factors-taking-risk-budgeting-one-step-further\/","title":{"rendered":"Risk factors: taking risk budgeting one step further"},"content":{"rendered":"

Many large investors already apply risk budgeting to asset
\nclasses in their allocation (although sometimes mistakenly
\nreferred to as \u2018risk parity\u2019, it is not always a question of
\nequally weighted risks). The main difficulty lies in defining
\nrisk budgets that are appropriate for each investor. By taking
\neconomic risk factors into account, we can accurately define
\nprecise risk budgets to match the specific objectives of each
\ninvestor profile. This represents a step up from the traditional
\nrisk parity approach, which focuses solely on asset classes.<\/strong><\/p>\n

The economic and financial crisis has prompted many
\npension funds and institutional investors to reconsider their
\nlong-term asset allocation strategy. They are calling into
\nquestion traditional portfolio optimisation as propounded
\nby Harry Markowitz and are increasingly leaning towards
\nthe risk budgeting approach and its corollary, risk parity.<\/p>\n

The principle underlying risk budgeting-based allocation
\nis simple: the asset allocation is established on the basis
\nof the risk contribution of each portfolio component,
\nrather than the expected return. In the case of a portfolio
\ncomprising two assets managed using the equallyweighted
\nrisks approach, each asset makes an equal
\ncontribution to risk and therefore to performance. To
\nachieve this balance, exposure to the riskier assets is
\nreduced, and vice versa.<\/p>\n

<\/p>\n

Traditional diversification via weightings leads to markedly
\ndifferent results. For example, in a balanced portfolio
\ncomposed of 50% equities and 50% bonds, the equity
\ncomponent accounts for almost 90% of the portfolio\u2019s
\nvolatility. By way of symmetry, the equity component will
\nalso generate the same proportion of performance.<\/p>\n

Risk-based diversification thus intuitively seems much
\nmore accurate and equitable. Investors can use it to
\noptimise their risk profile beyond mere diversification
\nbased on market capitalisation, and thus to obtain a better
\nrisk-adjusted return. In addition, it allows for an ex-ante
\nunderstanding of performance attribution: due to the
\nmirroring effect between risk and performance, investors
\ncan anticipate the source of their portfolio\u2019s performance.<\/p>\n

Risk factors<\/h2>\n

A major challenge remains to be addressed, namely
\ndefining a risk allocation that is in line with investors\u2019
\nobjectives. Even if a portfolio\u2019s allocation appears to be
\noptimally, or at least neutrally, diversified using the risk
\nparity approach, it may harbour other hidden sources
\nof risk, e.g. financial and economic factors affecting the
\nperformance of the asset classes in the portfolio. Insofar
\nas risk factors can affect more than one asset class at a
\ntime, ignoring them and focusing solely on asset classes
\ncan lead to a concentration of a limited number of factors.<\/p>\n

Long-term investors may go a step further in applying this
\napproach by no longer considering only asset classes
\n(equities, bonds, commodities, etc.) but also economic
\nrisk factors such as economic activity (GDP, industrial
\nproduction), inflation (commodity and consumer prices),
\ninterest rates (real interest rates, steepening and convexity
\nof the yield curve) and the effective exchange rate.<\/p>\n

Using such a framework, the portfolio is constructed on
\nthe basis of the risk budget allocated to these various
\neconomic criteria. At first glance, this might seem easily
\nachieved by simply linking an asset class with a particular
\ntype of risk \u2013 e.g. bonds with interest rate risk \u2013 and
\napplying the traditional risk budgeting method. However,
\nsuch an approach would not be effective as an asset
\nclass can be exposed to several economic risk factors at a
\ntime. While equities are affected by economic growth and
\nindustrial production, they are also impacted by interest
\nrates \u2013 as shown by the Gordon-Shapiro model \u2013 and
\ninflation. Therefore the sensitivity of each asset class to
\ncertain risk factors must be identified in order to establish
\nan asset allocation.<\/p>\n

It is all the more appropriate to consider economic risk
\nfactors rather than financial assets given that many pension
\nfunds and asset managers reason in macroeconomic
\nterms. For example, a pension fund expecting a lasting
\nperiod of growth may try to increase the GDP sensitivity
\nof its allocation. By assigning a budget per risk factor, it
\ncan do so with great accuracy by selecting a set of assets
\nsensitive to growth.<\/p>\n

Such an approach standardises the concept of risk parity in that several asset allocations can then be compared on the basis of common factors, regardless of the asset classes used. It also reconciles the quantitative approach to strategic asset allocation with the fundamental approach.<\/p>\n

Smart beta selection<\/h2>\n

Furthermore, an increasing number of pension funds are opting to invest in \u2018alternative\u2019 or \u2018smart beta\u2019 indices to supplement their passive management activities. Several competing methods currently exist, each with their own objectives. These include approaches such as the Equally Weighted Portfolio (EW), Minimum Variance Portfolio (MV), Equal Risk Contribution Portfolio (ERC) and the Most Diversified Portfolio (MDP), among others. Analysing the risk contribution of each factor by type of approach gives investors a clearer picture of the various competing smart beta methodologies.
\n\"ENG_-_Contribution_au_risque_des_facteurs.jpg\"<\/p>\n

By way of example, the S&P 100 index is sensitive to economic activity but it is also affected by interest rates as shown in the risk contribution table. This can be explained by the fact that the slope of the yield curve is a leading economic indicator. Interest rate risk is therefore not exclusive to bonds.<\/p>\n

<\/p>\n

Sensitivity to interest rate risk is higher in the case of minimum variance portfolios. As the equities making up an MV portfolio are generally among the least volatile in their universe, they have bond-like characteristics. Result: implementing the MV approach in the equity component introduces bond risk. From a strategic allocation viewpoint, this amounts to transferring part of the equity component allocation to the bond component.<\/p>\n

Analysing the risk contribution of each factor thus allows us to assess a portfolio\u2019s economic profile. It also provides a clear idea of sensitivity to economic risk factors and, above all, allows for a more accurate prediction of how \u201csmart beta\u201d indices will behave in different macroeconomic conditions.