2016년 3월 3일 목요일

[발췌: Haldane et al] What is the contribution of the financial sector (2010)

출처: Andrew Haldane, Simon Brennab, and Vasileios Madouros. "What is the contribution of the financial sector: Miracle or mirage?" Chapter 2 in The Future of Finance:  LSE Report. 2010


※ 발췌 (excerpt):
ABSTRACT: This chapter considers the contribution made by the financial sector to the wider economy. The measured GDP contribution of the financial sector suggests it underwent a "productivity miracle" from the 1980s onwards, as finance rose as share of national output despite a declining labour and capital share. But a detailed decomposition of returns to banking suggests an alternative interpretation: much of the growth reflected the effects of higher risk-taking. Leverage, higher trading profits and investments in deep-out-of-the money options were the risk-taking strategies generating excess returns to bank shareholders and staff. Subsequently, as these risks have materialised, returns to banking have reversed. In this sense, high pre-crisis returns to finance may have been more mirage than miracle. This suggests better measuring of risk-taking in finance is an important public policy objective─for statisticians and regulators, as well a for banks and their investors.


1. Introduction

  The financial crisis of the past three years has, on any measure, been extremely costly. As in past financial crises, public sector debt seems set to double relative to national income in a number of countries (Reinhart and Rogoff (2009)). And measures of foregone output, now and in the future, put the present value cost of the crisis at anywhere between one and five times annual world GDP (Haldane (2010)). Either way, the scars from the current crisis seem likely to be felt for a generation.

  It is against this backdrop that an intense debate is underway internationally about reform of finance (Goodhard (2010)). Many of the key planks of that debate are covered in other chapters in this volume. Some of these reform measures are extensions or elaborations of existing regulatory initiatives─for example, higher buffers of higher quality capital and liquidity. Others propose a reorientation of existing regulatory apparatus─for example, through counter-cyclical adjustments in prudential policy (Bank of England (2009b), Large (2010)). Others still suggest a root-and-branch restructuring of finance─for example, by limiting the size and/or scope of banking (Kay (2009), Kotlikoff(2010)).

  In evaluating these reform proposals, it is clearly important that the on-going benefits of finance are properly weighed alongside the cost of crisis. Doing so requires an understanding and measurement of the contribution made by the financial sector to economic well-being. This is important both for making sense of the past (during which time the role of finance has grown) and for shaping the future (during which it is possible the role of finance may shrink).

  ( ... ... ) Banking contributed to a Great Recession on a scale last seen at the time of the Great Depression.
  Yet the official statistics on the contribution of the financial sector paint a rather different picture. According the the National Accounts, the nominal gross value-added (GVA) of the financial sector in the UK grew at the fastest pace on record in 2008Q4. As a share of whole-economy output, the direct contribution of the UK financial sector rose to 9% in the last quarter of 2008. Financial corporations' gross operating surplus (GVA less compensation for employees and other taxes on production) increased b £5.0bn to £20bn, also the largest quarterly increase on record. At a time when people believed banks were contributing the least to the economy since the 1930s, the National Accounts indicated the financial sector was contributingg the most since the mid-1980s. How do we begin to square this cycle?

  ( ... ... ) In Section 3, we consider growth accountng breakdown ( ... ) This suggests banking has undergone, at least arithmetically, a "productivity miracle" over the past few decades. Section 4 explores in greater detail some of the quantitative drivers of high aggregate returns to banking, while Section 5 explores some of banks' business activities. Risk illusion, rather than a productivity miracle, appears to have driven high returns to finance. The recent history of banking appears to be as much mirage as miracle. ( ... )


2. Measuring Financial Sector Output


(a) Historical Trends in GVA


(b) Measuring GVA in the Financial Sector

  To begin to understand these trends, it is important first to assess how financial sector value-added is currently measured and the problems this poses when gauging the sector's contribution to the broader economy.

  ( ... ... )

  This is also the case for some of the services provided by the financial sector.[n.2]  For example, investment banks charge explicit fees when they advise clients on a merger or acquisition. Fees or commissionsare also levies on underwriting the issuance of securities and for the market-making activities undertaken for clients. But such direct charges account for only part of the financial systems's total revenues. Finance─and commercial banking in particular─relies heavily on interest flows as a means of payment for the services they provide. Banks charge an interest rate margin to capture these intermediation services.

  To measure the value of financial services embedded in interest rate margines, the concept of FISIM─Financial Intermediation Services Indirectly Measured─has been developed internationally. The concept itself was introduced in the 1993 update of the UN Systems of National Accounts (SNA). The SNA recognises that financial intemediaries provide services to consumers, businesses, governments and the rest of the world for which explicit charges are not made. In associate guidelines, a number of such services are identified including:
  • Taking, managing and transferring deposits;
  • Providing flexible payment mechanisms such as debit cards
  • Making loans or other investments; and
  • Offering financial advice or other business services.

  FISIM is estimated for loans and deposits only. The calculation is based on the difference between the effective rates of interest (payable and receivable) and a 'reference' rate of interest, multiplied by the stock of outstanding balances. According to SNA guidelines, ‘this reference rate represents the pur cost of borrowing funds─that is a rate from which the risk premium has been eliminated to the greatest extent possible, and that does not include any intermediation services.’[n.3] For example, a £1,000 loan with a 9% interest receivable and a 4% reference rate gives current price FISIM on the loan = £1,000×(9%-4%)=£50. And for a £1,000 deposit with a 3% interest payable and a 4% reference rate, this gives current price FISIM on the deposit =£1,000×(4%-3%)=£10. Overall, estimated current price FISIM accounts for a significant share of gross output of the banking sector (Chart 7).

  Estimating a real measure of FISIM is fraught with both conceptual and computational difficulties. In the earlier example of the second-hand car dealer, statisticians can use the the number of cars sold as an indicator of the volume of gross output. But the conceptual equivalent for financial intermediation is not clear. Would two loans of £50 each to the same customer represent a higher level of activity than one loan of £100? ( ... ... )


(c) Refining the Measurement of FISIM

  While the introduction of FISIM into the national accounts was an important step forward, it is not difficult to construct scenarios where the contribution of the financial sector to the economy could be mis-measured under this approach. A key issue is the extent to which bearing risk should be measured as a productive service provided by the banking system.

  (i) Adjusting FISIM for Risk

  Under current FISIM guidelines, which use risk-free policy rates to measure the reference rate, banks' compensation for bearing risk constitutes part of their measured nominal output. This can lead to some surprising outcomes. For example, assume there is an economy-wide increase in the expected level of defaults on loans or in liquidity risk, as occurred in October 2008. Banks will rationally respond by increasing interest rates to cover the rise in expected losses. FISIM will score this increased compensation for expected losses on lending as a rise in output. In other words, at times when risk is rising, the contribution of the financial sector to the real economy may be overestimated. This goes some way towars explaining the 2008Q4 National Accounts paradox of a rapidly rising financial sector contribution to nominal GDP.

  Of course, the financial sector does bear the risk of other agents in the economy. Banks take on maturity mismatch or liquidity risk on behalf of households and companies. And banks also make risky loans funded by debt, which exposes them to default or solvency risk. But it is not clear that bearing risk is, in itself, a productive activity. Any household or corporate investing in a risky asset is a fundamental feature of capital markets and is not specific to the activities of banks. Conceptually, therefore, it is not clear that risk-based income flows should represent bank output.

  The productive activity provided by an effectively functioning banking system might be better thought of as measuring and pricing credit and liquidity risk. For example, banks screen borrowers' creditworthiness when extending loans, thereby acting as delegated monitor. And they mananage liquidity risk through their treasury operations, thereby acting as delegated treasurer. These risk-pricing services are remunerated implicitly through interest rates banks charge to their customer.

  Stripping out the compensation for bearing risk to better reflect the service component of the financial sector could be achieved in different ways. One possibility would be to adjust FISIM using provisions as an indicator of expected losses. A broader adjustment for risk, as has been suggested by several commentators, would be to move away from the risk-free rate as the reference rate within FISIM.[n.4]  For example, a paper prepared for the OECD Working Party on National Accounts (Mink (2008)) suggested that the FISIM calculation should use reference rates that match the maturity and credit risk of loans and deposits. This would also eliminate an inconsistency within the current National Accounts framework. Measured financial intermediation output increases if a bank bears the risk of lending to a company. But gross output is unchanged if a household holds a bond issued by the same company and thus bears the same risk.

  ( ... ... )

(ii) Measuring risk

  ( ... ... )

  Attempting to adjust the measurement of bank output for risk by changing the reference rate in FISIM is an improvement on current practices. But it would still fall short of assessing whether the financial sector is pricing risk correctly and hence assessing the true value of the services banks provide to the wider economy. Unless the price of risk can be evaluated, it seems unlikely the contribution of the financial sector to the ecoomy can be measured with accuracy.


3. Decomposing the Contribution of the Financial Sector─the Productivity "Miracle"

  To that end, an alternative way of looking at the contribution of the financial sector is through inputs to the production process. ( ... ) Was this [rapid growth in finance] accompanied by a rising share of resources employed by finance relative to the rest of the economy? Or did it instead reflect unusually high returns to these factors of production? 

  ( ... ... )

  The time-series evidence is some respect even more dramatic. Philippon and REshelf (2009) have undertaken a careful study of "excess" wages in the US financial industry since the start of the previous century, relative to a benchmark wage. Chart 15 plots their measure of excess wages. This shows a dramatic spike upwards which commenced in the early 1980s, but which exploded from the 1990s onwards. The only equivalent wage spike was in the run-up to the Great Crash in 1929. Philippon and Reshelf attribute both of these wage spikes to financial deregulation.

  This picture is broadly mirrored when turning from returns to labour to returns to capital. In the 1950s gross profitability of the financial sector relative to capital employed was broadly in line with the rest of the economy (Chart 16). But since then, and in particular over the past decade, returns to capital have far outpaced those at an economy-wide level.

  Chart 17 plots UK banks' return on equity capital (RE) since 1920 (Alessadri and Haldane (2009)). Although conceptually a different measure of returns to capital, the broad message is the same. Trends in ROE are clearly divided into two periods. In the period up until around 1970, ROE in banking was around 7% with a low variance. In other words, returns to finance broadly mimicked those in the economy as a whole, in line with the gamble payoffs in Chart 6. But the 1970s mark a regime shift, with the ROE in banking roughly trebling to over 20%, again in line with gamble payoffs. Excess returns accumulated to captial as well as labour.

  These returns were by no means unique to UK banks. Chart 18 plots ROEs for major internationally active banks in the US and Europe during this century. Two features are striking. First, the level of ROEs was consistently at or above 20% and on a rising trend up until the crisis. This is roughly double ROEs in the non-financial sector over the period. Second, the degree of cross-country similarity in these ROE profiles is striking. This, too, is no coincidence. During much of this period, banks internationally were in engaged in a highly competitive ROE race. Therein lies part of the explanation for these high returns to labour and capital in banking.


4. Explaining Aggregate Returns in Banking ─ Excess Returns and Risk Illusion

  How do we explain these high, but temporary, excess returns to finance which appear to have driven the growing contribution of the financial sector to aggregate economic activity? In this section we discuss potential balance sheet strategies which may contributed to these rents. Essentially, high returns to finance have been driven banks assuming higher risks. Banks' profit, like their contribution to GDP, may have been flattered by the mis-measurement of risk.

  The crisis has subsequently exposed the extent of this increased risk-taking by banks. In particular, three (often related) balance sheet strategies for boosting risks and returns to banking were dominant in the run-up to crisis:

  • increased leverage, on and off-balance sheet;
  • increased share of asset held at fair value; and
  • writing deep out-of-the-money options.

  What each of these strategies had in common was that they generated a rise in balance sheet risk, as well as return. As importantly, this increase in risk was to some extent hidden by the opacity of accounting disclosure of the complexity of the products involved. This resulted in a divergence between reported and risk-adjusted returns. In other words, while reported ROEs rose, risk-adjusted ROEs did not (Haldane (2009)).

  To some extent, these strategies and their implications were captured to a degree in performance measures. For example, the rise in reported average ROEs of banks over the past few decades occurred alongside a rise in its variability. ( ... ... ) In that sense, the banking "productivity miracle" may have been, at least in part, a mirage─a simple, if dramatic, case of risk illusion by banks, investors and regulators.


(a) Increased leverage

  Banks' balance sheet have grown dramatically in relation to underlying economic activity over the past century. Charts 19 and 20 plot this ratio for the UK and the US over the past 130 years. For the US, there has been a secular rise in banks' assets from around 20% to over 100% of GDP. For the UK, a century of flat-lining at around 50% of GDP was broken in the early 1970s, since when banks' assets in relation to national income have risen tenfold to over 500% of GDP.

  This century has seen an intensification of this growth. According to data compiled by the ^Banker^, the balance sheets of the world's largest 1000 banks increased by around 150% between 2001 and 2009 (Chart 21). In cross-section terms, the scale of assets in the banking system now dwarfs that in other sectors. Looking at the size of the largest firm's assets in relation to GDP across a spectrum of industries, finance is by far the largest (Chart 22).

 The extent of balance sheet growth was, if anything, understated by banks' reported assets. Accounting and regulatory policies permitted banks to place certain exposures off-balance sheet, including special purpose vehicles and contingent credit commitments. Even disclosures of on-balance sheet positions on derivatives disguised some information about banks' contingent exposures.

  This rapid expansion of the balance sheet of the banking system was not accompanies by a commensurate increase in its equity base. Over the same 130 year period, the capital ratios of banks in the US and UK fell from around 15-25% at the start of the 20th century to around 5% at its end (Chart 23). In other words, on this metric measures of balance sheet leverage rose from around 4-times equity capital in the early part of the previous century to around 20 times capital at the end.

  If anything, the pressure to raise leverage increased further moving into this century. ( ... ... ) Among the major global banks in the world, levels of leverage were on average more than 50 times equity at the peak of the boom (Chart 25).

  ( ... ... )

  Taken together, this evidence suggests that much of the "productivity miracle" of high ROEs in banking appear to have been the result not of productivity gains on the underlying asset pool, but rather a simple leveraging up of the underlying equity in the business.


(b) Larger trading books


(c) Writing deep out-of-the-money options


 ( .. .... )

  What all of these strategies had in common was that they involved banks assuming risk in the hunt for yield─risk that was often disguised because it was parked in the tail of the return distribution. Excess returns─from leverage, trading books and out-of-the-money options─were built on an inability to measure and price risk. The productivity miracle was in fact a risk illusion. In that respect, mis-measurement of the contribution of banking in the National Accounts and the mis-measurement of returns to banking in their own accounts have a common underlying cause.

( ... ... )

댓글 쓰기