## _wp1638

## Source details

**Canonical URL:** [_wp1638](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1638.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1638.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1638.pdf.json)

---

### Procyclicality and the Case for Long-term Countercyclical Investment
- Willingness to bear risk tends to diminish in periods of stress and increase in upturns, amplifying market movements and creating feedback loops detrimental to financial stability and long-term economic growth.
- The Financial Stability Forum (2009, p. 8-10) and the Bank of England (BoE, 2014, p. 4) argue that procyclicality can decrease the resilience of the financial system and potentially interrupt vital financial functions.
- Industry critique (Focusing Capital on the Long-term Initiative, FCLTI, 2015, p. 3) highlights that short-term performance focus undermines corporate investment, holds back economic growth, and lowers returns for savers.
- The World Economic Forum (2011, p. 9) identifies three constituencies that benefit from greater emphasis on long-term countercyclical investment:
  - asset owners who might enjoy better returns by accessing risk premiums the rest of the market eschew;
  - corporations that can obtain financing for strategic initiatives with large upfront costs but significant long-term payoffs (i.e., infrastructure);
  - broader society by mitigating volatility and dislocations from boom-bust asset cycles.

### Short-termism, Herd Behavior, and Corporate Outcomes
- Many major players do not take a long-term approach, using short-term investment strategies that track benchmark indices and favor external managers who focus on short-term returns, producing herd behavior, excess volatility, bubbles, and suboptimal corporate decisions for long-term value creation (Barton and Wiseman, 2014).
- Definitions used:
  - “Countercyclical” investment: leaning against multi-year trends, possibly when asset valuations become stretched vis-à-vis historical norms.
  - “Long-term” investment: liabilities maturing more than a year in advance (with caveat that rebalancing, liquidity, and compliance can shorten actual holding periods).

### Under-attention to Institutional (Non-bank) Asset Owners in Policy Debates
- Policy focus has concentrated on dampening procyclicality in the banking sector (e.g., countercyclical capital buffers) while the role of procyclicality among institutional investors has received much less attention.
- Prior analyses have concentrated on asset managers rather than the underlying asset owners and consultants (Financial Stability Oversight Council, 2013; Feroli and others, 2014; IMF, 2015; Financial Stability Board, 2015).
- Papaioannou and others (2013, p. 5) note that “global and national financial sector regulation and supervision have yet to play a significant role in restraining such risks to the global financial system.”

### Importance and Scale of Asset Owners Relative to Banks and Asset Managers
- Asset owners oversee a pool of assets larger than banks; gap is widening.
- Only 25 to 35 percent of financial (excluding real estate) wealth worldwide is estimated to be managed via asset management firms—the majority is invested directly by asset owners themselves (McKinsey & Company, 2013; Jones, 2015; IMF, 2015).
- As of December 2013, assets held by the world’s largest 500 banks stood at US$97.1 trillion.
- Non-bank asset owners (pension funds, insurers, central banks, sovereign wealth funds, endowments, foundations and retail) oversee almost double this amount (relative to the $97.1 trillion in bank assets cited for the largest 500 banks).
- Since the global financial crisis, the stock of bank assets has grown at a considerably slower pace than that of invested financial assets (Jones, 2015).

### Distinction between Asset Owners, Consultants, and Asset Managers
- Much of the asset management industry provides implementation services: vehicles through which asset owners express objectives formulated by investment committees and board directives.
- Funds where asset managers have considerable discretion represent a minority of invested assets.
- In the open-end mutual fund industry (which oversees almost eighty percent of worldwide assets invested through collective investment schemes), asset managers typically have discretion only over tactical adjustments versus pre-set benchmarks; such tactical deviations are constrained by tracking error limits.
- It is “inappropriate and ineffective for asset managers to be viewed as responsible for actions that are essentially just the passing through of end-investor decisions” (Elliott, 2014, p. 1).
- Where asset managers amplify procyclicality through tactical decisions, this can reflect short-term performance appraisal terms imposed on them by asset owners and consultants.
- Distinguishing actions of asset owners and asset managers is important for policy formation aimed at reducing risks to economic and financial stability from procyclical investment behavior.

### Key Observations and Implications for Policy and Practice
- Long-term, countercyclical investment by large institutional investors can potentially:
  - improve returns for asset owners by accessing neglected risk premiums;
  - facilitate corporate investment in projects with large upfront costs and long paybacks;
  - reduce macro-financial volatility associated with boom-bust cycles.
- Regulatory and supervisory frameworks have not yet played a significant role in constraining procyclical risks posed by non-bank asset owners.
- Addressing procyclicality effectively requires attention to the incentives and governance of asset owners and their consultants, not only asset managers.

### Box 1 — Characteristics of Asset Owners, Consultants and Asset Managers
- Roles, legal status and functional responsibilities:
  - Asset Owners
    - Legal ownership over assets: Yes (Principal).
    - Legal responsibilities: Required to represent the interests of stakeholders pursuant with board directives and/or regulatory guidelines.
    - Key functional responsibility: Discretion over strategic asset allocation, proportion of assets managed internally vs. externally, and manager selection.
    - Solvency risk: Yes.
  - Consultants
    - Legal ownership over assets: No (Agent).
    - Legal responsibilities: Required to act as a fiduciary to clients.
    - Key functional responsibility: Advisory and due diligence services for asset owners on manager selection, risk management and strategic asset allocation.
    - Solvency risk: No.
  - Asset Managers
    - Legal ownership over assets: No (Agent).
    - Legal responsibilities: Required to act as a fiduciary to clients—investment decisions pursuant with guidelines in investment management agreement or fund constituent documents.
    - Key functional responsibility: Funds management on behalf of asset owners, usually tactical deviations vs. predetermined benchmarks.
    - Solvency risk: Not usually (aside from hedge funds and private equity).

- Representative market shares and concentration:
  - Just eight percent of assets under management (or $1.2 trillion) in the U.S. mutual fund industry are invested through ‘hybrid funds’ (a mix of stocks and bonds).
  - Hedge funds account for five percent of worldwide assets managed via collective investment vehicles; multi-asset, multi-region hedge funds collectively oversee $450 billion, around one fifth of hedge fund assets under management.

- Data coverage and asset owner sample (aggregate universe examined: $24.2 trillion in assets as of December 2014):
  - Global central banks: $6.6 trillion
  - U.S. private pension funds: $7.0 trillion
  - U.S. public (state and local) pension funds: $3.8 trillion
  - U.S. life insurers: $6.3 trillion
  - U.S. endowments and foundations: $0.5 trillion
  - Data sources and sample windows: Global central banks: US Treasury TIC System (June 1989—June 2014); IMF IFS for official gold holdings (June 1983—June 2014). Private and public U.S. pension funds, and U.S. life insurers: U.S. Federal Reserve Z.1 ‘Financial Accounts’ (December 1989—December 2014). U.S. Endowment Funds: NACUBO—Commonfund survey (June 1993—June 2014).

### Empirical approach and main findings on procyclicality and valuations
- Regression objective: Examine how year-to-year portfolio weight changes respond to contemporaneous relative returns, past relative returns over multiple horizons, and relative valuations (real yields).
- Regression variables included contemporaneous relative return (t : t-1), lagged relative return (t-1 : t-2), medium horizon relative returns (t-2 : t-5), long horizon relative returns (t-5 : t-10), and relative real yield at t-1 (real prospective yields defined as nominal yield less 10-year ahead Consensus Economics inflation forecast; equities use cyclically adjusted earnings yield).
- Main aggregate findings (almost a quarter century of annual data):
  - Asset allocation decisions of various types of asset owners appear procyclical at a multi-year horizon.
  - Procyclicality takes two forms of roughly equal importance:
    - Failure to rebalance in the current year (passive drift with contemporaneous relative returns).
    - Longer-term performance chasing (response to past relative returns).
  - Procyclical behavior is most evident for equities among asset classes.
- Summary statistics from regression results:
  - Relative returns: Of the 56 estimated coefficients associated with relative returns, 22 are statistically significantly positive (at higher than the 10 percent level), broadly evenly divided between contemporaneous and past year returns; 4 coefficients are statistically significant and negative.
  - Relative valuations: Of the 16 estimated coefficients for real yield differentials, 5 are positive (4 of which are statistically significant at higher than 10 percent), while 11 are negative (only 1 statistically significant).
  - Interpretation: Both passive drift and active return-chasing drive portfolio changes; countercyclical responses to relative valuations are not the dominant determinant of portfolio shifts for large asset owners.

- Asset-owner-specific procyclicality patterns:
  - Global central banks: strongest procyclicality for treasuries, equities and gold.
  - U.S. private pension funds, U.S. public pension funds, U.S. life insurers: strongest for fixed income and equities.
  - U.S. endowment funds: fixed income, equities and real estate.

- Model performance and robustness:
  - Across sixteen regressions, thirteen produce F-statistics statistically significant at higher than 5 percent.
  - Adjusted-R2’s range from an average of 0.41 for global central banks to an average of 0.79 for asset classes in U.S. life insurance portfolios.
  - Breusch-Godfrey serial correlation LM tests up to third order: null of no serial correlation could not be rejected at the 5 percent or higher level in any of the sixteen regressions.

### Proposed five-pillar framework to encourage long-term, countercyclical investment
- Framework elements (four industry-led, one regulatory):
  1. Strengthening governance
  2. Heterogeneous benchmarks, factor tilts and disciplined rebalancing
  3. Principal–agent alignment
  4. Risk management focused on long-term shortfall risk
  5. Regulatory conventions that avoid amplifying procyclicality

- Selected governance recommendations:
  - Consider minimum financial literacy accreditation standards for trustee appointments.
  - Emphasize accountability over implementation (not just articulation) of investment policy statements.
  - Preemptive stakeholder education to set expectations on interim mark-to-market losses when harvesting long-term risk premia.

- Benchmarks, factor tilts and rebalancing:
  - Critique: Capitalization-weighted benchmarks mechanically link market prices to index weights, inducing captive buying of securities with strongest price momentum and promoting system-wide herding.
  - Suggested alternatives and strategies:
    - Fundamental-weighted or value-weighted indices (countercyclical rebalancing toward risk premia or economic scale).
    - Harvesting elevated equity risk premia (timing tilt when equity risk premium is high).
    - Selling equity index volatility in high-volatility regimes.
    - Distressed value strategies that benefit long-horizon, volatility-agnostic investors.
  - Empirical illustrations referenced:
    - Fundamental index 10-year annualized outperformance assessment (10-year period ending June 2015).
    - U.S. Equity Risk Premium analysis over March 1962–December 2015 (214 quarterly observations).
    - Selling S&P500 put options: data June 1986–December 2015 (118 quarterly observations).
    - Distressed investment funds vs MSCI World: December 1989–December 2015 (312 monthly observations); Distressed Investment Funds (Ave return = 10.1%, St Dev = 6.5%, Sharpe = 1.08, Max Drawdown = 27%); MSCI World Equities (Ave return = 6.3%, St Dev = 15.6%, Sharpe = 0.21, Max Drawdown = 55%).

### Risk management: focus on symmetric shortfall risk
- Distinction by investor liability structure:
  - No leverage & long-duration liabilities (savings-based SWF, endowment): primary risk factor is long-term shortfall risk; marked-to-market frequency: Quarterly / Annual / Multi-year; symmetric concern for left and right tails.
  - Leverage & short-term liabilities (hedge funds): primary risk factor is short-term volatility; marked-to-market frequency: Daily; asymmetric focus on left tail.
- Core recommendation: Risk management for long-term asset owners should focus on minimizing long-term shortfall risk, recognizing that shortfall can arise from taking too little risk as well as too much risk.
- Illustration based on S&P500 (1962—2015):
  - Constant buy-and-hold average annual excess total return: 4.6%
  - Excluding best quintile of returns: -0.2% (compound excess annual return)
  - Excluding worst quintile of returns: 9.4% (compound excess annual return)
  - Differential vs baseline in both cases: 4.8 percentage points per annum (symmetry between incurring large losses and foregoing large gains).

### Principal–agent frictions and manager compensation
- Procyclical hiring and firing:
  - Hired managers had positive excess returns prior to hiring and fired managers had negative excess returns prior to firing; post-termination, fired managers tended to perform at least as well as those that replaced them (Goyal and Wahal, 2008).
  - Career risk and relative ranking incentives can induce agents to herd and prefer momentum strategies, amplifying systemic procyclicality.
- Manager compensation asymmetries:
  - Fees based on unrealized capital gains produce call option–like payoffs (managers retain upside but not downside), incentivizing procyclical risk taking and bubble participation.
  - Fund-level features that mitigate redemption risk (e.g., closed-end structures) can better support valuation-based, market-stabilizing styles, but closed-end funds comprised just 1 percent of assets in collective investment vehicles as of 2014 (open-end mutual funds comprised 41 percent).

### Policy and industry implications (selected)
- Where investment behavior creates negative externalities, policy intervention may be warranted; industry-led remedial actions (governance, benchmarks, incentives, risk frameworks) are a viable alternative and largely feasible.
- Practical steps include minimum trustee accreditation, heterogeneous benchmarks aligned with liabilities and risk premia, long-term aligned compensation structures, symmetric shortfall-focused risk frameworks, and regulatory conventions that avoid exacerbating procyclicality.

### The Fee-Based Wedge: paying performance fees on unrealized gains
- Scenario parameters:
  - Asset returns of 7.5% in years 1-4; a return of -25% in year 5.
  - Annual performance fee of 20%.
  - Hurdle rate on performance of 1%.
  - Annual management fee of 1 %.
  - No upfront sales load, and no tracking error.
  - Fees extracted on the basis of year-end portfolio values.
- Key quantitative finding:
  - After 5 years when investor sells and the market is unchanged, the investor has paid out a cumulative 12.8 % in fees and is substantially worse off despite unchanged underlying market value.
- Conceptual implication:
  - Performance fees paid on unrealized gains create a wedge between the market value of the portfolio (unchanged) and the investor’s net outcome (worsened by cumulative fees), particularly if a large negative return realizes later.

### Principal—agent frictions and remedies (selected)
- Behavioral/incentive problems:
  - Managers may earn performance fees on interim gains yet receive no compensation for subsequent large losses (example: 7.5 percent per annum for years 1—4 and -25 percent in final year).
  - High watermarks and short-term appraisals can induce managers to behave procyclically or to close and reopen funds to reset compensation prospects.
- Recommended adjustments to align interests:
  - De-emphasize recent performance in external manager appraisals and adopt longer appraisal windows (e.g., 5–10 years rather than 1–3 years).
  - Use closed-end funds and lockup arrangements; employ tiered exit fee structures and defer portions of performance-based cash payments while a longer-term track record builds.
  - Adopt symmetrical fee structures or multi-year clawback provisions.
  - Encourage managers to commit a meaningful portion of their own wealth to the funds they manage.
  - Base hiring, firing, and compensation on broader value-added activities.
  - Advocate for more in-house management of capital where capacity building is realistic.

### Mitigating procyclical regulation and accounting effects
- Problem statement:
  - Fair value (‘mark-to-market’) accounting and risk-sensitive capital/funding requirements can induce procyclical selling in stressed markets when regulatory metrics force asset reallocation.
- Empirical notes:
  - Defined benefit pension plans experienced significant swings in funding ratios from surplus to deficit around the crisis (pre-crisis maxima in September 2007/October 2008; post-crisis minima in March 2009).
  - U.S. insurance companies more capital constrained by fair value losses sold RMBS at much lower prices through the crisis.
- Policy concepts to reduce regulatory-induced procyclicality:
  - Avoid automatic de-risking in response to a shift from surplus to deficit; instead incorporate expectations of time-varying forward returns in glide-path remedial plans.
  - Require funds comfortably in surplus to lock-in funding status by matching liabilities with assets (derisk).
  - Report funding status using spot (real time), smoothed (moving average), and constant discount rate assumptions.
  - Make credit rating changes trigger internal review rather than involuntary investment actions; consider a simple rating corridor with different thresholds for asset allocation changes.
  - Provide clarity on circumstances where plan sponsors can use assets other than cash to temporarily cover funding deficits and define eligible asset types.

### Examples of ad hoc cross-jurisdictional actions (illustrative)
- Pension funds:
  - Extension of solvency recovery plan: Canada (2008), UK (2009), Netherlands (2009), Ireland (2012)
  - Flexibility/smoothing in discount rates used to value liabilities: Denmark (2008), Netherlands (2010)
  - Changes to benefits: Switzerland (2008)
  - Increased range of assets eligible to be used in place of sponsor cash contributions: UK (2006-)
  - Changes to funding requirements: Finland (2008), Japan (2009), Canada (2010)
- Insurers:
  - Changes to solvency requirements: UK (2001-04; 2008-09), US (2007-09), Switzerland (2013)
  - Changes to valuation methods: UK (2001-04), US (2007-09)
  - Changes to discount rates: Sweden (2001-12), UK (2001-04), Denmark (2008, 2012), Netherlands (2012)
  - Extension of solvency recovery plan: Sweden (2011)

### Transparency and stabilizing investment: framework principles
- Central insight:
  - Transparency via fair value accounting need not force procyclical investment if regulatory frameworks contain inbuilt stabilizers and avoid mechanically requiring immediate remedial actions based on backward-looking metrics.
- Recommended features of a countercyclical regulatory framework:
  - Accumulate resilience in good times so constraints can be relaxed in stress periods via automatic stabilizers rather than discretionary ad hoc measures.
  - Ensure regulatory responses are timely, symmetric, and transparent to reduce uncertainty and moral hazard.

### Conclusion and avenues for future research
- Summary policy message:
  - Patient, countercyclical investment by long-term asset owners can improve individual returns and contribute to systemic stability; achieving this requires changes in governance, benchmarks, principal—agent arrangements, risk management calibration, and regulatory conventions.
- Suggested research directions:
  - Construct longer and higher-frequency time series of individual asset owner portfolios to assess evolution of procyclicality.
  - Cross-country comparisons of institutional investment behavior.
  - Deeper analysis of optimal incentive structures across the investment chain, including within asset owners and the role of consultants.
  - Evaluate initiatives to stimulate direct investment from long-term asset owners (e.g., capital relief, government guarantees for greenfield infrastructure).
  - More granular appraisal of institutional and regulatory impediments to market-stabilizing long-term investment across different classes of asset owners.

### Annex 1 — Descriptive Statistics for Annual Changes in Asset Allocation Weights (selected excerpts)
- Global Central Banks — Annual changes in allocation to U.S. Treasuries, U.S. Agencies, U.S. Credit, U.S. Equities, Gold
  - Average: -0.5%, 0.2%, 0.1%, 0.3%, -0.3%
  - Median: -0.6%, 0.5%, 0.0%, 0.4%, -0.3%
  - Standard Deviation: 3.3%, 3.0%, 0.5%, 1.6%, 1.0%
  - Maximum: 8.9%, 5.5%, 1.1%, 2.7%, 1.8%
  - Minimum: -5.2%, -6.7%, -1.0%, -3.7%, -1.9%
  - Skew: 0.85, 0.03, 0.03, -0.74, 0.28
  - Kurtosis: 3.87, 2.78, 3.56, 3.24, 2.53
  - Jarque-Bera: 3.82, 1.21, 0.33, 2.35, 0.72
  - P-value: 0.15, 0.55, 0.85, 0.31, 0.70
  - Total Observations: 25, 25, 25, 25, 32
  - Start of Sample: Jun-89, Jun-89, Jun-89, Jun-89, Jun-82
  - End of Sample: Jun-14, Jun-14, Jun-14, Jun-14, Jun-14

- Annual Changes in Asset Allocation Weights (three paired series)
  - Fixed Income, Equities series:
    - Average: -0.2%, 0.6% | -1.0%, 1.1% | -1.0%, 0.9%
    - Median: -0.1%, 1.3% | -1.2%, 0.5% | -1.5%, 1.5%
    - Standard Deviation: 2.9%, 3.5% | 3.3%, 3.6% | 2.5%, 2.7%
    - Maximum: 8.7%, 5.7% | 8.2%, 7.5% | 5.9%, 5.0%
    - Minimum: -4.3%, -10.4% | -8.2%, -9.4% | -4.5%, -7.4%
    - Skew: 0.99, -1.30 | 0.36, -0.55 | 1.10, -1.31
    - Kurtosis: 4.41, 5.26 | 4.59, 4.41 | 4.03, 4.66
    - Jarque-Bera: 6.18, 12.30 | 3.16, 3.34 | 6.14, 10.07
    - P-value: 0.05, 0.00 | 0.21, 0.19 | 0.05, 0.01
    - Total Observations: 25, 25 | 25, 25 | 25, 25
    - Start of Sample: Dec-89, Dec-89 | Dec-89, Dec-89 | Dec-89, Dec-89
    - End of Sample: Dec-14, Dec-14 | Dec-14, Dec-14 | Dec-14, Dec-14

- U.S. Private Pensions, U.S. Public Pensions, U.S. Life Insurers, U.S. Endowments — selected asset classes
  - Average: -0.9%, -0.1% | 0.3%, 0.5% | 0.1%
  - Median: -1.3%, 0.0% | 0.2%, 0.5% | 0.0%
  - Standard Deviation: 1.7%, 2.7% | 0.5%, 0.8% | 0.4%
  - Maximum: 2.0%, 4.0% | 1.9%, 2.3% | 0.8%
  - Minimum: -3.8%, -5.9% | -0.4%, -0.8% | -1.0%
  - Skew: 0.39, -0.59 | 1.54, 0.17 | -0.58
  - Kurtosis: 2.15, 2.82 | 5.11, 2.96 | 4.74
  - Jarque-Bera: 1.17, 1.26 | 12.23, 0.10 | 3.83
  - P-value: 0.56, 0.53 | 0.00, 0.95 | 0.15
  - Total Observations: 21, 21 | 21, 21 | 21
  - Start of Sample: Jun-93, Jun-93 | Jun-93, Jun-93 | Jun-93
  - End of Sample: Jun-14, Jun-14 | Jun-14, Jun-14 | Jun-14

- Notes on model fit figures:
  - Global Central Banks (Table 2): Overall R² = 0.68; Adj R² = 0.59. U.S. Treasuries: R² = 0.62; Adj R² = 0.51. U.S. Agencies: R² = 0.29; Adj R² = 0.11. U.S. Credit: R² = 0.56; Adj R² = 0.44. U.S. Equities: R² = 0.48; Adj R² = 0.38.
  - U.S. Private Pension Funds, U.S. Public Pension Funds, U.S. Life Insurers (Table 3): Private Pensions (Fixed Income): R² = 0.43; Adj R² = 0.28. Private Pensions (Equities): R² = 0.65; Adj R² = 0.56. Public Pensions (Fixed Income): R² = 0.48; Adj R² = 0.34. Public Pensions (Equities): R² = 0.73; Adj R² = 0.66. Life Insurers (Fixed Income): R² = 0.77; Adj R² = 0.71. Life Insurers (Equities): R² = 0.89; Adj R² = 0.86.
  - U.S. Endowment Funds (Table 4): Overall R² = 0.68; Adj R² = 0.58. Endowment Funds (Fixed Income): R² = 0.86; Adj R² = 0.81. Endowment Funds (Equities): R² = 0.28; Adj R² = 0.04. Endowment Funds (Private Equity): R² = 0.39; Adj R² = 0.18. Endowment Funds (Hedge Funds): R² = 0.63; Adj R² = 0.51.

*Source: _wp1638 - References; Box 1; Annex 1 content as provided in the source PDF.*

### References .............................................................................................................

### _wp1638 - References .............................................................................................................

### Procyclicality and the Case for Long-term Countercyclical Investment
- Willingness to bear risk tends to diminish in periods of stress and increase in upturns, which can amplify market movements and create feedback loops detrimental to financial stability and long-term economic growth.
- The Financial Stability Forum (2009, p. 8-10) and the Bank of England (BoE, 2014, p. 4) argue that procyclicality can decrease the resilience of the financial system and potentially interrupt vital financial functions.
- Industry critique of “quarterly capitalism” (Focusing Capital on the Long-term Initiative, FCLTI, 2015, p. 3) highlights how short-term performance focus undermines corporate investment, holds back economic growth, and lowers returns for savers.
- The World Economic Forum (2011, p. 9) identifies three constituencies that benefit from greater emphasis on long-term countercyclical investment:
  - asset owners who might enjoy better returns by accessing risk premiums the rest of the market eschew;
  - corporations that can obtain financing for strategic initiatives with large upfront costs but significant long-term payoffs (i.e., infrastructure);
  - broader society by mitigating volatility and dislocations from boom-bust asset cycles.

### Short-termism, Herd Behavior, and Corporate Outcomes
- Barton and Wiseman (2014) argue that many major players do not take a long-term approach, using short-term investment strategies that track benchmark indices and favor external managers who focus on short-term returns, which produces herd behavior, excess volatility, bubbles, and suboptimal corporate decisions for long-term value creation.
- Definitions used in the paper:
  - “Countercyclical” investment: leaning against multi-year trends, possibly when asset valuations become stretched vis-à-vis historical norms.
  - “Long-term” investment: liabilities maturing more than a year in advance (with caveat that rebalancing, liquidity, and compliance can shorten actual holding periods).

### Under-attention to Institutional (Non-bank) Asset Owners in Policy Debates
- Policy focus has concentrated on dampening procyclicality in the banking sector (e.g., countercyclical capital buffers) while the role of procyclicality among institutional investors has received much less attention.
- Prior analyses have concentrated on asset managers rather than the underlying asset owners and consultants (cited documents include Financial Stability Oversight Council, 2013; Feroli and others, 2014; IMF, 2015; Financial Stability Board, 2015).
- Papaioannou and others (2013, p. 5) note that “global and national financial sector regulation and supervision have yet to play a significant role in restraining such risks to the global financial system.”

### Importance and Scale of Asset Owners Relative to Banks and Asset Managers
- Asset owners oversee a pool of assets larger than banks; gap is widening.
- Only 25 to 35 percent of financial (excluding real estate) wealth worldwide is estimated to be managed via asset management firms—the majority is invested directly by asset owners themselves (McKinsey & Company, 2013; Jones, 2015; IMF, 2015).
- As of December 2013, assets held by the world’s largest 500 banks stood at US$97.1 trillion.
- Non-bank asset owners (pension funds, insurers, central banks, sovereign wealth funds, endowments, foundations and retail) oversee almost double this amount (relative to the $97.1 trillion in bank assets cited for the largest 500 banks).
- Since the global financial crisis, the stock of bank assets has grown at a considerably slower pace than that of invested financial assets (Jones, 2015).

### Distinction between Asset Owners, Consultants, and Asset Managers
- Much of the asset management industry provides implementation services: vehicles through which asset owners express objectives formulated by investment committees and board directives.
- Funds where asset managers have considerable discretion over allocations to asset classes and geographies represent a minority of invested assets.
- In the open-end mutual fund industry (which oversees almost eighty percent of worldwide assets invested through collective investment schemes), asset managers typically have discretion only over tactical adjustments versus pre-set benchmarks pertaining to a single asset class and region; such tactical deviations are constrained by tracking error limits.
- It is “inappropriate and ineffective for asset managers to be viewed as responsible for actions that are essentially just the passing through of end-investor decisions” (Elliott, 2014, p. 1).
- Where asset managers amplify procyclicality through tactical decisions, this can often reflect the short-term performance appraisal terms imposed on them by asset owners and consultants.
  - Asset owners and consultants have imperfect knowledge of manager skill in a high-noise, low-signal environment, which can lead to terminating managers after periods of underperformance and thereby favoring managers holding securities with strong momentum.
- Distinguishing actions of asset owners and asset managers is important for policy formation aimed at reducing risks to economic and financial stability from procyclical investment behavior.

### Key Observations and Implications for Policy and Practice
- Long-term, countercyclical investment by large institutional investors can potentially:
  - improve returns for asset owners by accessing neglected risk premiums;
  - facilitate corporate investment in projects with large upfront costs and long paybacks;
  - reduce macro-financial volatility associated with boom-bust cycles.
- Regulatory and supervisory frameworks have not yet played a significant role in constraining procyclical risks posed by non-bank asset owners.
- Addressing procyclicality effectively requires attention to the incentives and governance of asset owners and their consultants, not only asset managers.

*Source: _wp1638 - References .............................................................................................................*

### Box 1. Characteristics of Asset Owners, Consultants and Asset Managers

### Box 1. Characteristics of Asset Owners, Consultants and Asset Managers

### Roles, legal status and functional responsibilities
- Asset Owners
  - Legal ownership over assets: Yes (Principal).
  - Legal responsibilities: Required to represent the interests of stakeholders pursuant with board directives and/or regulatory guidelines.
  - Key functional responsibility: Discretion over strategic asset allocation, proportion of assets managed internally vs. externally, and manager selection.
  - Solvency risk: Yes.
- Consultants
  - Legal ownership over assets: No (Agent).
  - Legal responsibilities: Required to act as a fiduciary to clients.
  - Key functional responsibility: Advisory and due diligence services for asset owners on manager selection, risk management and strategic asset allocation.
  - Solvency risk: No.
- Asset Managers
  - Legal ownership over assets: No (Agent).
  - Legal responsibilities: Required to act as a fiduciary to clients—investment decisions pursuant with guidelines in investment management agreement or fund constituent documents.
  - Key functional responsibility: Funds management on behalf of asset owners, usually tactical deviations vs. predetermined benchmarks.
  - Solvency risk: Not usually (aside from hedge funds and private equity).

### Representative market shares and concentration
- Just eight percent of assets under management (or $1.2 trillion) in the U.S. mutual fund industry are invested through ‘hybrid funds’ (a mix of stocks and bonds).
- Hedge funds account for five percent of worldwide assets managed via collective investment vehicles; multi-asset, multi-region hedge funds collectively oversee $450 billion, around one fifth of hedge fund assets under management.

### Data coverage and asset owner sample
- Aggregate universe examined: $24.2 trillion in assets (as of December 2014), comprising:
  - Global central banks: $6.6 trillion
  - U.S. private pension funds: $7.0 trillion
  - U.S. public (state and local) pension funds: $3.8 trillion
  - U.S. life insurers: $6.3 trillion
  - U.S. endowments and foundations: $0.5 trillion
- Data sources and sample windows:
  - Global central banks: US Treasury TIC System (June 1989—June 2014); IMF IFS for official gold holdings (June 1983—June 2014).
  - Private and public U.S. pension funds, and U.S. life insurers: U.S. Federal Reserve Z.1 ‘Financial Accounts’ (December 1989—December 2014).
  - U.S. Endowment Funds: NACUBO—Commonfund survey (June 1993—June 2014).

### Empirical question and regression approach
- Objective: Examine how year-to-year portfolio weight changes respond to contemporaneous relative returns, past relative returns over multiple horizons, and relative valuations (real yields).
- Time-series regression (annual data) for each asset class and asset owner where:
  - Change in portfolio weight between year t and t-1 is regressed on:
    - Contemporaneous relative return (t : t-1)
    - Relative return lagged one year (t-1 : t-2)
    - Relative returns over medium horizon (t-2 : t-5)
    - Relative returns over long horizon (t-5 : t-10)
    - Relative real yield at t-1 (real prospective yields defined as nominal yield less 10-year ahead Consensus Economics inflation forecast; equities use cyclically adjusted earnings yield)

### Key empirical findings on procyclicality and valuations
- Main aggregate findings based on almost a quarter century of annual data:
  - Asset allocation decisions of various types of asset owners appear procyclical at a multi-year horizon.
  - Procyclicality takes two forms of roughly equal importance:
    - Failure to rebalance in the current year (passive drift with contemporaneous relative returns).
    - Longer-term performance chasing (response to past relative returns).
  - Procyclical behavior is most evident for equities among asset classes.
- Summary statistics from regression results:
  - Relative returns: Of the 56 estimated coefficients associated with relative returns, 22 are statistically significantly positive (at higher than the 10 percent level), broadly evenly divided between contemporaneous and past year returns; 4 coefficients are statistically significant and negative.
  - Relative valuations: Of the 16 estimated coefficients for real yield differentials, 5 are positive (4 of which are statistically significant at higher than 10 percent), while 11 are negative (only 1 statistically significant).
  - Interpretation: Both passive drift and active return-chasing drive portfolio changes; countercyclical responses to relative valuations are not the dominant determinant of portfolio shifts for large asset owners.
- Asset-owner-specific procyclicality patterns:
  - Global central banks: strongest procyclicality for treasuries, equities and gold.
  - U.S. private pension funds, U.S. public pension funds, U.S. life insurers: strongest for fixed income and equities.
  - U.S. endowment funds: fixed income, equities and real estate.
- Model performance and robustness:
  - Across sixteen regressions (various asset classes and investor types), thirteen produce F-statistics statistically significant at higher than 5 percent.
  - Adjusted-R2’s range from an average of 0.41 for global central banks to an average of 0.79 for asset classes in U.S. life insurance portfolios.
  - Breusch-Godfrey serial correlation LM tests up to third order: null of no serial correlation could not be rejected at the 5 percent or higher level in any of the sixteen regressions.

### Proposed five-pillar framework to encourage long-term, countercyclical investment
- High-level framework elements (four industry-led, one regulatory):
  1. Strengthening governance
  2. Heterogeneous benchmarks, factor tilts and disciplined rebalancing
  3. Principal–agent alignment
  4. Risk management focused on long-term shortfall risk
  5. Regulatory conventions that avoid amplifying procyclicality

### Governance recommendations (selected measures)
- Consider minimum financial literacy accreditation standards for trustee appointments.
- Emphasize accountability over implementation (not just articulation) of investment policy statements.
- Preemptive stakeholder education to set expectations on interim mark-to-market losses when harvesting long-term risk premia.

### Benchmarks, factor tilts and rebalancing
- Critique of prevailing practice:
  - Capitalization-weighted benchmarks mechanically link market prices to index weights, inducing captive buying of securities with strongest price momentum and promoting system-wide herding.
- Suggested market-stabilizing alternatives and strategies:
  - Fundamental-weighted or value-weighted indices (countercyclical rebalancing toward risk premia or economic scale).
  - Harvesting elevated equity risk premia (timing tilt when equity risk premium is high).
  - Selling equity index volatility in high-volatility regimes.
  - Distressed value strategies that benefit long-horizon, volatility-agnostic investors.
- Empirical illustration references in source:
  - Fundamental index 10-year annualized outperformance assessment (10-year period ending June 2015).
  - U.S. Equity Risk Premium analysis over March 1962–December 2015 (214 quarterly observations).
  - Selling S&P500 put options: data June 1986–December 2015 (118 quarterly observations).
  - Distressed investment funds vs MSCI World: December 1989–December 2015 (312 monthly observations); Distressed Investment Funds (Ave return = 10.1%, St Dev = 6.5%, Sharpe = 1.08, Max Drawdown = 27%); MSCI World Equities (Ave return = 6.3%, St Dev = 15.6%, Sharpe = 0.21, Max Drawdown = 55%).

### Risk management: focus on symmetric shortfall risk
- Distinction by investor liability structure:
  - No leverage & long-duration liabilities (savings-based SWF, endowment): primary risk factor is long-term shortfall risk; marked-to-market frequency: Quarterly / Annual / Multi-year; symmetric concern for left and right tails.
  - Leverage & short-term liabilities (hedge funds): primary risk factor is short-term volatility; marked-to-market frequency: Daily; asymmetric focus on left tail.
- Core recommendation: Risk management for long-term asset owners should focus on minimizing long-term shortfall risk, recognizing that shortfall can arise from taking too little risk as well as too much risk.
- Illustration: Based on S&P500 (1962—2015)
  - Constant buy-and-hold average annual excess total return: 4.6%
  - Excluding best quintile of returns: -0.2% (compound excess annual return)
  - Excluding worst quintile of returns: 9.4% (compound excess annual return)
  - Differential vs baseline in both cases: 4.8 percentage points per annum (symmetry between incurring large losses and foregoing large gains).

### Principal–agent frictions and manager compensation
- Procyclical hiring and firing:
  - Empirical patterns documented that hired managers had positive excess returns prior to hiring and fired managers had negative excess returns prior to firing; post-termination, fired managers tended to perform at least as well as those that replaced them (Goyal and Wahal, 2008).
  - Career risk and relative ranking incentives can induce agents to herd and prefer momentum strategies, amplifying systemic procyclicality.
- Manager compensation asymmetries:
  - Fees based on unrealized capital gains produce call option–like payoffs (managers retain upside but not downside), incentivizing procyclical risk taking and bubble participation.
  - Fund-level features that mitigate redemption risk (e.g., closed-end structures) can better support valuation-based, market-stabilizing styles, but closed-end funds comprised just 1 percent of assets in collective investment vehicles as of 2014 (open-end mutual funds comprised 41 percent).

### Policy and industry implications (selected)
- Where investment behavior creates negative externalities, policy intervention may be warranted; industry-led remedial actions (governance, benchmarks, incentives, risk frameworks) are a viable alternative and largely feasible.
- Practical steps include minimum trustee accreditation, heterogeneous benchmarks aligned with liabilities and risk premia, long-term aligned compensation structures, symmetric shortfall-focused risk frameworks, and regulatory conventions that avoid exacerbating procyclicality.

*Source: Author; Box title and content as provided in the source PDF._wp1638 - Box 1. Characteristics of Asset Owners, Consultants and Asset Managers*

### 7.5 percent per annum from years 1—4, and a negative return of 25 percent in the final year,

### _wp1638 - 7.5 percent per annum from years 1—4, and a negative return of 25 percent in the final year,

### The Fee-Based Wedge: paying performance fees on unrealized gains
- Scenario parameters (as reported in the source notes):
  - Asset returns of 7.5% in years 1-4; a return of -25% in year 5.
  - Annual performance fee of 20%.
  - Hurdle rate on performance of 1%.
  - Annual management fee of 1 %.
  - No upfront sales load, and no tracking error.
  - Fees extracted on the basis of year-end portfolio values.
- Key quantitative finding:
  - After 5 years when investor sells and the market is unchanged, the investor has paid out a cumulative 12.8 % in fees and is substantially worse off despite unchanged underlying market value.
- Conceptual implication:
  - Performance fees paid on unrealized gains create a wedge between the market value of the portfolio (unchanged) and the investor’s net outcome (worsened by cumulative fees), particularly if a large negative return realizes later.

### Principal—agent frictions and remedies
- Behavioral and incentive problems:
  - Managers may earn performance fees on interim gains yet receive no compensation for subsequent large losses (example: 7.5% per annum for years 1—4 and -25% in final year).
  - High watermarks and short-term appraisals can induce managers to behave procyclically or to close and reopen funds to reset compensation prospects.
- Recommended adjustments to align interests:
  - De-emphasize recent performance in external manager appraisals and adopt longer appraisal windows (e.g., 5–10 years rather than 1–3 years) to better distinguish skill from luck.
  - Use closed-end funds and lockup arrangements to reduce short-term redemption incentives; employ tiered exit fee structures and defer portions of performance-based cash payments while a longer-term track record builds.
  - Adopt symmetrical fee structures (manager exposed to downside) or multi-year clawback provisions to shift compensation from a call option–like payoff to a forward-like payoff.
  - Encourage managers to commit a meaningful portion of their own wealth to the funds they manage to better align personal interests with principals.
  - Base hiring, firing, and compensation on broader value-added activities (network building, staff training) given difficulty in predicting future excess returns.
  - Advocate for more in-house management of capital where capacity building is realistic, accounting for holistic cost—benefit tradeoffs.

### Mitigating procyclical regulation and accounting effects
- Problem statement:
  - Fair value (‘mark-to-market’) accounting and risk-sensitive capital/funding requirements can induce procyclical selling in stressed markets when regulatory metrics force asset reallocation.
- Empirical notes:
  - Defined benefit pension plans experienced significant swings in funding ratios from surplus to deficit around the crisis (Figure 10; pre-crisis maxima in September 2007/October 2008; post-crisis minima in March 2009).
  - U.S. insurance companies more capital constrained by fair value losses sold RMBS at much lower prices through the crisis.
- Policy concepts to reduce regulatory-induced procyclicality:
  - Avoid automatic de-risking in response to a shift from surplus to deficit; instead incorporate expectations of time-varying forward returns in glide-path remedial plans.
  - Require funds comfortably in surplus to lock-in funding status by matching liabilities with assets (derisk).
  - Report funding status using spot (real time), smoothed (moving average), and constant discount rate assumptions to provide clarity and flexibility.
  - Make credit rating changes trigger internal review rather than involuntary investment actions; consider a simple rating corridor with different thresholds for asset allocation changes.
  - Provide clarity on circumstances where plan sponsors can use assets other than cash to temporarily cover funding deficits and define eligible asset types.

### Examples of ad hoc measures (illustrative cross-jurisdictional actions)
- Pension funds:
  - Extension of solvency recovery plan: Canada (2008), UK (2009), Netherlands (2009), Ireland (2012)
  - Flexibility/smoothing in discount rates used to value liabilities: Denmark (2008), Netherlands (2010)
  - Changes to benefits: Switzerland (2008)
  - Increased range of assets eligible to be used in place of sponsor cash contributions: UK (2006-)
  - Changes to funding requirements: Finland (2008), Japan (2009), Canada (2010)
- Insurers:
  - Changes to solvency requirements: UK (2001-04; 2008-09), US (2007-09), Switzerland (2013)
  - Changes to valuation methods: UK (2001-04), US (2007-09)
  - Changes to discount rates: Sweden (2001-12), UK (2001-04), Denmark (2008, 2012), Netherlands (2012)
  - Extension of solvency recovery plan: Sweden (2011)

### Transparency and stabilizing investment: framework principles
- Central insight:
  - Transparency via fair value accounting need not force procyclical investment if regulatory frameworks contain inbuilt stabilizers and avoid mechanically requiring immediate remedial actions based on backward-looking metrics.
- Recommended features of a countercyclical regulatory framework:
  - Accumulate resilience in good times so constraints can be relaxed in stress periods via automatic stabilizers rather than discretionary ad hoc measures.
  - Ensure regulatory responses are timely, symmetric, and transparent to reduce uncertainty and moral hazard.

### Conclusion and avenues for future research
- Summary policy message:
  - Patient, countercyclical investment by long-term asset owners can improve individual returns and contribute to systemic stability; achieving this requires changes in governance, benchmarks, principal—agent arrangements, risk management calibration, and regulatory conventions.
- Suggested empirical and policy research directions:
  - Construct longer and higher-frequency time series of individual asset owner portfolios to assess evolution of procyclicality.
  - Cross-country comparisons of institutional investment behavior.
  - Deeper analysis of optimal incentive structures across the investment chain, including within asset owners and the role of consultants.
  - Evaluate initiatives to stimulate direct investment from long-term asset owners (e.g., capital relief, government guarantees for greenfield infrastructure).
  - More granular appraisal of institutional and regulatory impediments to market-stabilizing long-term investment across different classes of asset owners.

*Source: Author.*

### Annex 1. Descriptive Statistics for Annual Changes in Asset Allocation Weights

### Annex 1. Descriptive Statistics for Annual Changes in Asset Allocation Weights

### Notes
- See Section II for a description of the underlying data series for each class of asset owner.

### Global Central Banks — Annual changes in allocation to U.S. Treasuries, U.S. Agencies, U.S. Credit, U.S. Equities, Gold
- Average: -0.5%, 0.2%, 0.1%, 0.3%, -0.3%
- Median: -0.6%, 0.5%, 0.0%, 0.4%, -0.3%
- Standard Deviation: 3.3%, 3.0%, 0.5%, 1.6%, 1.0%
- Maximum: 8.9%, 5.5%, 1.1%, 2.7%, 1.8%
- Minimum: -5.2%, -6.7%, -1.0%, -3.7%, -1.9%
- Skew: 0.85, 0.03, 0.03, -0.74, 0.28
- Kurtosis: 3.87, 2.78, 3.56, 3.24, 2.53
- Jarque-Bera: 3.82, 1.21, 0.33, 2.35, 0.72
- P-value: 0.15, 0.55, 0.85, 0.31, 0.70
- Total Observations: 25, 25, 25, 25, 32
- Start of Sample: Jun-89, Jun-89, Jun-89, Jun-89, Jun-82
- End of Sample: Jun-14, Jun-14, Jun-14, Jun-14, Jun-14

### Annual Changes in Asset Allocation Weights (three paired series shown in source)
- Annual Changes in Asset Allocation Weights for: Fixed Income, Equities | Fixed Income, Equities | Fixed Income, Equities
  - Average: -0.2%, 0.6% | -1.0%, 1.1% | -1.0%, 0.9%
  - Median: -0.1%, 1.3% | -1.2%, 0.5% | -1.5%, 1.5%
  - Standard Deviation: 2.9%, 3.5% | 3.3%, 3.6% | 2.5%, 2.7%
  - Maximum: 8.7%, 5.7% | 8.2%, 7.5% | 5.9%, 5.0%
  - Minimum: -4.3%, -10.4% | -8.2%, -9.4% | -4.5%, -7.4%
  - Skew: 0.99, -1.30 | 0.36, -0.55 | 1.10, -1.31
  - Kurtosis: 4.41, 5.26 | 4.59, 4.41 | 4.03, 4.66
  - Jarque-Bera: 6.18, 12.30 | 3.16, 3.34 | 6.14, 10.07
  - P-value: 0.05, 0.00 | 0.21, 0.19 | 0.05, 0.01
  - Total Observations: 25, 25 | 25, 25 | 25, 25
  - Start of Sample: Dec-89, Dec-89 | Dec-89, Dec-89 | Dec-89, Dec-89
  - End of Sample: Dec-14, Dec-14 | Dec-14, Dec-14 | Dec-14, Dec-14

### U.S. Private Pensions, U.S. Public Pensions, U.S. Life Insurers, U.S. Endowments — selected asset classes
- Annual Changes in Asset Allocation Weights for: U.S. Private Pensions (Fixed Income, Equities), U.S. Public Pensions (Fixed Income, Equities), U.S. Life Insurers (Private Equity, Hedge Funds, Real Estate)
  - Average: -0.9%, -0.1% | 0.3%, 0.5% | 0.1%
  - Median: -1.3%, 0.0% | 0.2%, 0.5% | 0.0%
  - Standard Deviation: 1.7%, 2.7% | 0.5%, 0.8% | 0.4%
  - Maximum: 2.0%, 4.0% | 1.9%, 2.3% | 0.8%
  - Minimum: -3.8%, -5.9% | -0.4%, -0.8% | -1.0%
  - Skew: 0.39, -0.59 | 1.54, 0.17 | -0.58
  - Kurtosis: 2.15, 2.82 | 5.11, 2.96 | 4.74
  - Jarque-Bera: 1.17, 1.26 | 12.23, 0.10 | 3.83
  - P-value: 0.56, 0.53 | 0.00, 0.95 | 0.15
  - Total Observations: 21, 21 | 21, 21 | 21
  - Start of Sample: Jun-93, Jun-93 | Jun-93, Jun-93 | Jun-93
  - End of Sample: Jun-14, Jun-14 | Jun-14, Jun-14 | Jun-14

### Notes on model fit figures (Annex 2 referenced in the source)
- For Global Central Banks (model coefficients and standard errors presented in Table 2):
  - Overall R² = 0.68; Adj R² = 0.59
  - U.S. Treasuries: R² = 0.62; Adj R² = 0.51
  - U.S. Agencies: R² = 0.29; Adj R² = 0.11
  - U.S. Credit: R² = 0.56; Adj R² = 0.44
  - U.S. Equities: R² = 0.48; Adj R² = 0.38
  - Gold: (figure shown; no separate R² listed beyond overall)
- For U.S. Private Pension Funds, U.S. Public Pension Funds, U.S. Life Insurers (model coefficients and standard errors presented in Table 3):
  - Private Pensions (Fixed Income): R² = 0.43; Adj R² = 0.28
  - Private Pensions (Equities): R² = 0.65; Adj R² = 0.56
  - Public Pensions (Fixed Income): R² = 0.48; Adj R² = 0.34
  - Public Pensions (Equities): R² = 0.73; Adj R² = 0.66
  - Life Insurers (Fixed Income): R² = 0.77; Adj R² = 0.71
  - Life Insurers (Equities): R² = 0.89; Adj R² = 0.86
- For U.S. Endowment Funds (model coefficients and standard errors presented in Table 4):
  - Overall R² = 0.68; Adj R² = 0.58
  - Endowment Funds (Fixed Income): R² = 0.86; Adj R² = 0.81
  - Endowment Funds (Equities): R² = 0.28; Adj R² = 0.04
  - Endowment Funds (Private Equity): R² = 0.39; Adj R² = 0.18
  - Endowment Funds (Hedge Funds): R² = 0.63; Adj R² = 0.51
  - Endowment Funds (Real Estate): (figure shown; no separate R² listed beyond above)

*Source: Annex 1, “Descriptive Statistics for Annual Changes in Asset Allocation Weights,” _wp1638 - Annex 1. Descriptive Statistics for Annual Changes in Asset Allocation Weights.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp1638.pdf_