COST TO INCOME RATIOS IN AUSTRALASIAN BANKING

Transcription

1 COST TO INCOME RATIOS IN AUSTRALASIAN BANKING DAVID TRIPE CENTRE FOR BANKING STUDIES MASSEY UNIVERSITY Contact details: Address: David Tripe Senior Lecturer Department of Finance Banking and Property Massey University Private Bag Palmerston North New Zealand Phone: ext 2337 Fax: Abstract: This paper looks at the cost to income ratio, also known as the efficiency ratio or expense to income ratio, and examines its usefulness as a measure of bank performance. The paper identifies ways in which the ratio can reflect factors such as differing bank structures, rather than necessarily being a measure of bank performance. The latter part of the paper explores conceptual relationships between cost to income ratios and profitability and considers whether superior cost to income performance is actually achievable. 1

2 Cost to income ratios seem to be the rage in banking, and Australasian banks are no exception. Banks compare themselves with their peers, and bank managements impress on their staff the need to reduce cost to income ratios to international best practice. Staff are told that costs must be reduced to achieve this, and although the sentiment is not usually specifically expressed, the suggested solution is to get rid of unnecessary staff and make existing staff work harder. Market analysts also look at the ratios reported, and make predictions about the future performance of the banks that report them. These comments are presumed to have an impact on bank share prices, another way in which the performance of bank managements are measured. 1 This paper focuses on cost to income ratios in Australasian banking. It looks at the way in which banks focus on this ratio, and at what it actually measures. In doing so it seeks to identify some of the weaknesses in using cost to income ratios as a basis for bank management decisions. This entails some review of the ways in which focus on cost to income ratios (instead of other measures of bank performance) can lead an observer to draw wrong conclusions in respect of a bank s performance. The obverse of this is to identify ways in which banks can manipulate their cost to income ratios so as to make them appear more satisfactory. Finally, the paper tries to make some judgements on the usefulness of the cost to income ratio for analysing bank performance. 1 Data to support the arguments raised in the paper are derived mainly from the author s ongoing analyses of the performance of Australian and New Zealand banks. Quarterly data is now available for New Zealand banks, reflecting the new disclosure regime for registered banks introduced at the beginning of Satisfactory Australian information is generally only available on an annual basis, from published annual reports in respect of the banks global operations. 2

3 THE FOCUS ON THE COST TO INCOME RATIO The Australasian bank best known for its focus on the cost to income ratio is the National Australia Bank. As long ago as its 1993 Annual Report, bringing the cost ratios of the offshore banks closer to the figure achieved in the Australian bank [p 9] was identified as a key objective. Each subsequent year s report has focused on cost to income ratio in the discussion of group financial highlights. The National Australia Bank is not alone, however. Cost to income ratios have also been important in the ANZ Bank in recent years, and complaints about the high cost to income ratio were believed to have been a significant factor behind the departure of former Managing Director Don Mercer. In the announcement of first half results for 1998, considerable emphasis has been placed on the reduction achieved in cost to income ratios. Westpac have also used the cost to income ratio as a basis for discussing the bank s costs, and the ratio was the subject of comment by the Chairman in his report to shareholders in the 1995 Annual Report (p 3). Westpac seems to place less emphasis on the ratio than do some other banks, however. 2 St George Bank focuses on its cost to income ratio in its 1997 Annual Report, amongst its highlights (pp 10-11) and Adelaide Bank also comments on its expense to income ratio amongst the highlights of its performance (p 3 of the 1997 Annual Report). It is not only the banks that focus on cost to income ratios. Commentators such as 2 In its 1992 Annual Report (p 6), the Bank announced its intention to achieve a cost to income ratio of 58% by By 1995, the ratio had only been reduced to 60.7%, and the failure to achieve the promised result may be a reason for the relative silence. 3

4 the financial press, consultancy firms and sharebroking analysts all look at cost to income ratios, sometimes subjecting them to further analysis. An example of this is in a recent Australian banking review, where reported cost to income ratios were used as a measure of banks achievements in reducing costs [Buttle (1998)]. Nor is it only in the Australian market that there is a focus on cost to income ratios. Recent research by Price Waterhouse found that the single key driver that has most impact in enhancing shareholder value remains reduction of the cost to income ratio [Madewell (1997), p 26]. Beyond that, a recent report by McKinsey & Co consultants has concluded that banks with cost to income ratios above 55 to 60% are under threat of takeover, unless the higher costs are a short term consequence of an earlier acquisition [James et al (1997)]. 3 WHAT IS THE COST TO INCOME RATIO? The standard definition of the cost to income ratio is as non-interest costs, excluding bad and doubtful debt expense, divided by the total of net interest income and non-interest income. Although the ratio is dependent on figures for both cost and income, its use tends to focus on costs. Non-interest costs, are perceived as that part of a bank s costs which are most controllable, and most responsive to management action. A reduction in costs, for a fixed level of revenue, should lead to increased profit, and thus increased return on equity and share price, the measures of greatest interest to investors in bank shares. Focusing on banks non-interest costs means that fluctuations in the general level of interest rates do not cause the volatility in the ratio that would arise if interest costs were included. Likewise, using net rather than total interest income in calculating the ratio reduces the 3 It is interesting to see that this argument has been taken up by the popular press. See Greising et al (1998), p 33. 4

5 volatility that fluctuations in the general level of interest rates would otherwise cause to income as the ratio s denominator. It may be argued that net interest income is influenced by the general level of interest rates, but net interest income is less prone to fluctuation, and thus less likely to cause undue variation in the cost to income ratio. 4 The cost to income ratio does not include bad and doubtful debt expense. The rationale for this is that such expense generally reflects the quality of credit decisions made in earlier periods, rather than the current performance of the bank. Moreover, if doubtful debt expense were included in the cost to income ratio, the ratio would be distorted when major write-offs were undertaken. 5 The effect of bad and doubtful debts and impaired assets is not totally eliminated in reviewing banks cost to income ratios, however. Other things being equal, higher levels of impaired assets and provisions would be expected to be accompanied by higher levels of operating costs and lower levels of income (as a smaller proportion of assets are generating income). 6 COST TO INCOME VS COST TO ASSETS To assess the usefulness and appropriateness of the cost to income ratio, however, we need to compare it to other measures which focus on banks costs. Another common ratio is 4 There are two points to be made here: (a) Fluctuations in the general level of interest rates may still influence the cost to income ratio as a result of the effect of the level of interest rates on economic activity - see further discussion below - and (b) The standard a priori assumption is that bank net interest margins tend to be higher when the general level of interest rates was rising or relatively high, and vice versa for lower margins. The author does not yet have sufficient data over a long enough time period to confirm this. 5 Such as when Australasian banks acknowledged the problems in their property-related lending in the late 1980s and early 1990s. 6 Other things being equal, therefore, a bank with a better quality loan book should have a lower cost to income ratio. 5

6 operating costs to average total assets, which, for comparative purposes, is most appropriately presented on an annualised basis. Although this ratio focuses on cost, it can be criticised. 7 The ratio of cost to average assets is affected by a bank s business mix. A bank with a lot of corporate/wholesale lending or placements in the inter-bank market should have a lower cost to assets ratio than a bank with a predominantly retail lending book. On the other hand, an investment bank that focuses on deal-making rather than running a large lending book would be expected to have a relatively high cost to assets ratio. 8 It can thus be argued that the cost to assets ratio is more a reflection of a bank s business mix rather than a basis for detailed analysis of its costs. There is also a problem with cross-border comparisons of cost to average assets ratios, in that banks in different countries may face different cost structures (and may have their financial accounts prepared in different ways). These variations are probably more important in their effect on the cost to assets ratio than on the cost to income ratio, as is shown by the figures for a selection of countries for 1995, in Table 1. 9 TABLE 1 - COMPARATIVE COST RATIOS Cost to assets ratio Cost to income ratio Australia 2.60% 64.7% Canada 2.50% 52.3% France 1.41% 65.6% Germany 1.64% 64.1% Iceland 4.85% 72.6% New Zealand 2.98% 66.5% 7 Some of these criticisms also apply to the cost to income ratio. 8 An example of such a bank in Australasia would be Macquarie Bank, which for the year to 31 March 1998 showed a cost to assets ratio of 7.09%. Cost to income ratio was 75.0%. 9 Data has been derived from OECD (1997), augmented by the author s calculations. 6

7 Switzerland 1.73% 56.4% United Kingdom 2.60% 63.9% United States 3.65% 63.4% It is interesting to note the close similarity between the cost to assets ratios for banks in Australia, Canada, and the United Kingdom, with the ratio for New Zealand only a little higher. This presumably reflects similar structures in the underlying banking systems, with the market dominated by a relatively small number of nationally operating banks. Cultural factors also mean that the types of business undertaken are relatively similar, while all countries have a relatively high proportion of cheques in their payment system volume. 10 The relatively low cost to assets ratios for the French, German and Swiss banks appear to reflect the structure of bank balance sheets in those countries. Interbank exposures are a relatively high proportion of both assets and liabilities, while banks also have significant security holdings. In the case of German banks, there is significant funding from bonds. By contrast with cost to assets, the cost to income ratio shows less variation, and might thus be seen as a better basis for bank performance assessment. Use of cost to income in preference to cost to assets can also be justified on the basis of the relationship between the cost to income ratio and bank profitability. Profitability is by definition the difference between income and cost, and a ratio that includes both of these must lead to a focus on profitability. The relationship between cost to income ratio and profitability is not direct, however, as the ratio requires a division whereas a profitability calculation requires a subtraction. This is one of the ways in which the ratio has the potential to mislead analysts. SOME PROBLEMS WITH THE COST TO INCOME RATIO The previous section of this paper looked at the cost to income ratio, and identified why it can be superior to cost to average assets for bank performance analysis. We now look at the 10 Bekier & Nickless (1998) argue that differences in the use of cheques in different countries payment systems is a major contributor to differences in the costs of those countries banking systems. 7

8 ratio more closely, to uncover ways in which it can be manipulated, and lead to misleading or even incorrect conclusions. The cost to income ratio is a ratio: it is thus affected not only by banks costs, but also by variations in incomes. For any given level of cost relative to a bank s assets, a reduction in income will cause an increase in the cost to income ratio. This reduction in income might be a reflection of a bank s ineffectuality in generating income (in which case the cost to income ratio is appropriately identifying inefficiency), but it may also be a reflection of a change in competitive conditions reducing the margins available to banks. A downturn in income might also reflect an economic downturn, reducing banks opportunities to undertake profitable business from which to earn interest and fees. This effect has already been observed in comparing cost to income ratios between banks in t Australia and New Zealand [Tripe (1996)]. If we look at the operations of individual banks, comparing gross income from their New Zealand operations and their global operations, income levels in New Zealand are lower. It follows from the differences in income levels that, even if cost levels were the same, there should be expected to be differences in the cost to income ratios for the banks different operations. In fact, costs relative to assets in New Zealand are lower that for the banks global operations, but cost to income ratios are lower for banks global operations than for their New Zealand operations. This has an unfortunate negative effect for managements of New Zealand banks. Their Australian bosses may be telling them that they are failing to achieve the desired group cost to income ratio, and yet they are already achieving lower cost to assets ratios than their Australian counterparts. This is a particular issue for the National Australia Bank. The cost to income ratio for the latest year for the Australian bank was 55.9% [1997 Annual Report, p 13], significantly below the global figure, calculated by the author at 57.6%. This highlights another difficulty with cost to income ratios, in that calculations by bank managements are not always transparent relative to published financial information. 11 The National Australia Bank s 1997 Annual Report identifies a group cost to income ratio of 11 Refer also Harris (1996), p 16. 8

9 55.9% [p 2], whereas the figure calculated directly from the income statement is 57.6%. The Bank s Annual Report states that the cost to income ratio for the bank s New Zealand operations was 56.5%, as opposed to the author s calculations from the published figures of 55.8%. 12 The discrepancy in respect of the global figures appears to relate to the treatment of a goodwill write-down in its operating expenses. This introduces general problems with accounting policy, which in some cases may be perceived as bank manipulation of accounting information to achieve desired results. 13 There was an interesting case with the Bank of New Zealand. Detailed examination of the bank s statement of financial performance in the 31 December 1996 General Short Form Disclosure Statement showed operating expenses for the quarter of $130 million. If these figures had been used, unadjusted, the cost to income ratio for that quarter would have been reduced to 48.7%. In fact, however, these operating expenses were after a write back of an excess provision against litigation of $22 million, and if this was classified as an extraordinary item, the cost to income ratio attained a more normal level of 56.9%. 14 In any case, differences in accounting treatments have the potential to distort the cost to income ratios banks report, or which are able to be calculated from their financial statements. 15 THE EFFECTS OF DIFFERENT STRUCTURES 12 The Bank of New Zealand s Annual Review for the year to 30 September 1997 also identifies the cost to income ratio as 55.8% (p 1). 13 There is an interesting analysis of the ANZ Bank s 1997 annual results in Stock (1998). 14 These accounting issues may be the cause of the discrepancy in the figures quoted for cost to income ratios. 15 An example of the way the ratio could be manipulated would be in alternative accounting treatments that could be applied where a bank uses mortgage brokers. Fees paid by customers are typically passed on in full or part to the broker, for their remuneration. A bank could either take the whole fee to income and expense the payment to the broker, or take to income only the net fees received. Where net fees are taken to income, the cost to income ratio will 9

10 Another concern with cost to income ratios is that differences in banks financial structures and patterns of business will cause differences in the ratios reported. This can be of significance in respect of the ways banks raise deposits, their mix of business, and in respect of excess levels of capital. The effect of different approaches to deposit raising is demonstrated by the Cheltenham & Gloucester Building Society, prior to its acquisition by Lloyds TSB [Davis (1994), p 28; Welch (1994), p 25; Harris (1996), p 11]. This institution specialised in raising deposits by mail, which meant that it faced much lower branch network costs. 16 A corresponding example in New Zealand is BNZ Finance, which has essentially no branch network. Its cost to income ratios for recent periods have been around 30%, substantially lower than for any of the major banks operating in New Zealand with a retail network. A bank that cuts costs in its branch network can afford to increase its interest rates so as to attract deposits, improving its cost to income ratio without affecting the bank s profitability. This can be demonstrated in a hypothetical scenario. Suppose we have a bank with $1 billion in assets, on which it is earning 10% gross interest income, but no non-interest income. At the outset it is having to pay 7% for its funds ($900 million) which it raises through a branch network which costs it $15 million a year. Overheads related to its Head Office are $10 million a year. A summary income statement for this bank is set out in the left-hand column of Table 2. TABLE 2 - THE EFFECT OF CHANGING THE MODE OF COLLECTING DEPOSITS $M Initial Position Higher interest scenario be lower than if gross fees are taken to income and payments to brokers treated as expenses. 16 More generally, Welch identifies the much lower cost to income ratios enjoyed by building societies in the United Kingdom, compared to banks, which he attributes to differences in product markets and strategies. 10

11 Interest income Interest expense Net interest income Operating expenses Net profit before tax Cost to income ratio 67.6% 57.1% Suppose the bank now decides to cut back on its branch network, reducing its annual network operating cost from $15 million to $6 million. To compensate for the reduced level of deposits through branches the bank raises deposit interest rates by 1%, leaving the aggregate level of deposits unaltered. As the scenario in the right-hand column of Table 2 shows, the bank s net profit is unaltered, but the cost to income ratio is significantly improved. In fact, an increase in interest rates could have a much more positive effect on the bank s depositraising, allowing for an even greater reduction in operating costs, although it is likely that the operating cost reduction would be tempered by increased advertising expenditure. 17 In New Zealand the ANZ Banking Group has consistently shown a relatively high cost to income ratio. This is not because of any deficiency in income-generation: historically the ANZ has been the most successful of the New Zealand banks in generating income, with the highest level of gross income. This can be largely attributed to its particular success in earning non-interest income, levels of which have been much higher than for its competitors. 18 Competitive pressures limited the ANZ s ability to earn atypical levels of income from interest margin. 17 The bank might also face increased liquidity risk as a result of losing its lower cost retail deposits. 18 Another example of differences in non-interest income can be seen by comparing essentially similar functions undertaken by banks in different countries, with quite different cost and revenue implications. Compare initiating a residential mortgage in the United States with doing so in Australia (or New Zealand). In the United States, the process is significantly more costly, but significantly more revenue is generated [Neagle (1995)]. Depending on the exact relative levels of cost and income, this would be expected to cause differences in bank cost to income ratios in the two markets. 11

12 The ANZ was thus a high-income, high cost bank, but the author s analysis shows that this was significantly a consequence of the accounting policies followed. A change in the accounting policy at the March 1998 half year has resulted in the separate itemisation of depreciation on UDC s leased assets, which has allowed an different approach to analysis of the information, which means in turn that gross income and cost to assets can now be seen as more directly comparable with those of the ANZ s competitors. The cost to income ratio is now nearer the norm for New Zealand banks, although still a little higher. One might expect the ANZ to show as a higher cost bank in any case. Australian experience demonstrates that purely domestic banks have lower costs than the banks with international operations. Banks with significant international networks and international operations, and involvement in non-retail banking, are inclined to have higher costs, but this will also be reflected in higher levels of income relative to assets. Another relationship between bank structure and cost to income ratios arises in respect of banks with excess capital. Holding excess capital allows a bank to undertake additional wholesale lending or investment activity at very low cost, increasing gross income without any corresponding increase in operating expenses. This is the same factor that underlies the relatively low cost to assets ratios of some of the European banks identified in Table 1. This situation may be better explained using an example. Suppose we have a bank with capital of $100 million, total assets of $1200 million generating an average income of 10%, but risk-weighted assets of only $800 million. The ratio of capital to risk-weighted assets is thus 12.5%, a typical level for Australasian banks. The average cost of funds for this hypothetical bank is 7%, and annual operating costs are $25 million. A summary income statement for the bank is shown in the left-hand column of Table 3. TABLE 3 - THE EFFECT OF EXCESS CAPITAL $M Initial position Position with use made of excess capital Interest income $120 $158 Interest expense $77 $113 12

13 Net interest income $43 $45 Operating expenses $25 $25 Net profit before tax $18 $20 Cost to income ratio 58.14% 55.56% Return on assets (before tax) 1.5% 1.25% The bank decides to use its excess capital by undertaking additional wholesale lending which, despite a low margin, still boosts aggregate profitability. It thus writes wholesale assets (to private corporations, subject to risk-asset weighting of 100%) of $400 million, at 9.5%, which it funds from wholesale deposits at 9%. The ratio of capital to risk-weighted assets is now 8.33%. The impact on the bank s operating costs of this additional lending is negligible. An income statement reflecting the changed position is shown in the right-hand column of Table 3. The bank has reduced its cost to income ratio, by an amount likely to be regarded as highly satisfactory by bank management. This reduction in cost to income ratio is at the expense of a significant reduction in return on assets, however, and, by utilising the excess capital to put on low-yield assets, the riskiness of the bank has been increased. 19 The bank s return on equity has increased marginally, but it is not clear that the reduction in the cost to income ratio will have made the bank s shareholders better off. A way in which this effect could be further accentuated would be if some of the loan assets acquired were high risk and high margin, such as for commercial real estate. Although entailing higher levels of overhead cost than for pure wholesale lending, margins are wider, and a bank is able to increase its net interest income by relatively much more than its noninterest costs. The higher risk of the loans may lead to a higher rate of credit losses, but these will not be directly reflected in the cost to income ratio [Toevs & Zizka (1994)]. Higher levels of equity can help a bank improve its cost to income ratio in another way. Equity is not only a regulatory requirement, but also a source of funds. As a source of funds, equity involves less administrative cost than do deposits [Wall (1983), p 46]. Other things 19 No allowance has been made for credit risk in respect of the new assets acquired, although there is obviously an assumption that they are effectively risk free. 13

14 being equal, therefore, a bank with more equity will have a lower cost to income ratio. Balance sheet structural differences thus show that it is not necessarily valid to compare cost to income ratios for banks in different countries. Yet bank analysts do so, such as is seen with the National Australia Bank s desire to have all member banks in the group with the same cost to income ratio as in Australia, and to reduce the ratio for the Australian bank to the level achieved by one or two exceptional banks worldwide. The belief in a pure international standard to which all banks should aspire is, in the author s view, unrealistic. 14

15 COST TO INCOME AND PROFIT The preceding discussion has suggested a relationship, albeit imperfect, between a bank s cost to income ratio and its profit. Osborne (1995) has given this relationship a specific algebraic formulation as follows: Cost to income ratio = 1 - {[ROE/(1 - tax rate)] X Equity/Income} - Bad debt expense/income Based on tests for United States banks over the period 1989 to 1993, Osborne concludes that the cost to income ratio is only weakly correlated with the individual elements of this equation. He therefore argues that one should not expect a stable relationship between a bank s return on equity and its cost to income ratio. Even without looking at correlations, the above equation might suggest that any relationship between cost to income ratio and return on equity would not be strong, because of the range of other variables involved. A decrease in cost to income ratio would only be reflected in an increased return on equity if all other elements on the right-hand side of the equation remained more or less constant. In fact, however, the individual elements on the right-hand side of the equation may not be as individually volatile as Osborne suggests. General experience in the New Zealand market is that banks pay near enough to the standard tax rate, and actual bank tax rates do not show much variation. Similarly the changes in economic conditions that underlie banks bad debt expense tend to have similar effects on all banks in the market, unless some bank has suffered some unusually bad luck with its borrowers. 20 If we therefore assume that the tax rate and rate of bad debt expense to income are more or 20 One would expect greater variations in the ratio of bad debt expense to income in the United States because banks there have greater concentrations of business in particular regional markets, which will be affected unevenly by economic cycles (and where economic problems which would give rise to increased bad debt expense have often been region- 15

16 less independent, in that they apply to conditions in the banking sector as a whole, Osborne s equation can be reduced to a formula along the following lines: Cost to income ratio = Ω [(ROE X Equity)/Income] where Ω is a function incorporating the other elements in the original equation, and which will imply a negative relationship (so that an increase in ROE would reflect a reduced cost to income ratio). This revised formula almost becomes a truism. If a bank s equity changes in relative terms, return on equity is likely to change in the opposite direction, and the multiple of the two will not necessarily change by very much at all. To further explore the relationship between the cost to income ratio and return on equity, some statistical analysis has been undertaken in respect of banks operating in New Zealand. A series of regressions has been run with return on equity as dependent variable and a range of individual ratios as independent variables, to ascertain the R square values applying to each of the independent variables studied for each bank. Ratios used as independent variables were bad debt expense to average total assets, cost to income ratio, gross income to average total assets, net interest income to average total assets, non-interest income to average total assets, and operating expenses to average total assets. Results are reported in Table 4. TABLE 4 - INFLUENCES ON RETURN ON EQUITY ANZ ASB BNZ CWB NBNZ TSB Bad Debt expense Cost to income Gross income Net interest income Non-interest income Operating expenses It is interesting to note that, in four cases out of six, cost to income ratio generates the highest specific, such as in Texas and New England in the late 1980s). 16

17 R square value, although the margin between the different ratios, in terms of their impact, is not always large. It is noted in the case of the ANZ Bank that, although the cost to income ratio provides the strongest relationship, this relationship is still relatively weak. Although there appears to be a relationship between cost to income ratio and return on equity, it is not evident that this relationship is strong. It would be unwise to conclude from this data in respect of New Zealand banks that a change in cost to income ratio should, in general, have an impact on any particular bank s return on equity. SOME CONSEQUENCES OF A FOCUS ON COST TO INCOME RATIO Insofar as a focus on cost to income ratio leads banks to review their costs, looking at ways to reduce them, concentration on the ratio may not be a bad thing. The danger, however, is that the cost to income ratio may become an end in itself for bank managers, promoting slash and burn cost-cutting, not in the best interests of long-term profit maximisation, and which does not thereby add value for bank shareholders. 21 It should be remembered that the cost to income ratio is only a ratio, and a bank s cost to income ratio may have changed because of changes to any one of a number of underlying factors. If the cost to income ratio is to be a driving force in bank managerial decision-making, it would be supposed that it should remain stable throughout the economic cycle, but there is no evidence that this is the case. A number of the elements that make up the cost and income used in the calculation of the ratio are sensitive to changes in economic conditions, and an examination of these allows assessment of the possible effect of economic conditions on cost to income ratios, as set out in Table 5. It is thus to be expected that a bank s cost to income ratio should increase in economic 21 Davis (1994) suggests that a long-term but continuous approach to cost-cutting is what is necessary to run a successful low-cost bank (p 27). 17

18 downturns, and reduce in improving economic conditions. The author is unconvinced that he yet has access to sufficient satisfactory data to confirm the validity of this supposition, but some assessment can be attempted. TABLE 5 - WHAT HAPPENS TO THE COST TO INCOME RATIO? Income/cost category Interest income Interest expense Net interest income Other income Gross income Network operating expenses Credit risk management expenses Total operating expenses Cost to income ratio Effect of economic downturn Decline with reduction in lending Decline due to less vigorous pursuit of deposits Net decline, as effect on income likely to be greater than effect on expense Decline in lending fees and transaction-related commissions Significant decline Some reduction as result of reduction in wage pressure Should be expected to increase Probably neutral overall, although possibility of slight reduction Increase For example, one can look at the trend in cost to income for the National Australia Bank group as a whole. The Australian economy has generally reflected improving economic conditions over the period in question, and yet there has not been much movement in the cost to income ratio, even though it has been targeted by management for reduction. Declining interest margins appear to be why reduced operating expenses have not lead to a corresponding reduction in the cost to income ratio. Efforts to reduce costs have only allowed the cost to income ratio to be held stable. As a further test, the author has looked at trends in Australian banks cost to income ratios over a longer period of time, as shown in Figure Figures have been calculated from OECD (1997), p 29. The high cost to income ratio in 1992 is likely to reflect the high levels of doubtful debts the banks had at that time (Refer to the previous discussion of this issue). 18

19 FIGURE 1 COST TO INCOME RATIOS - AUSTRALIA 75% 70% 65% 60% 55% 50% The cost to income ratio for all banks has remained relatively stable, despite all the banks concerted efforts at cost cutting. There was an upward movement to the ratio, away from its normal level, in 1992, which may be assumed to reflect the impact of the acknowledgement of bad debts in that year by the major banks. There is thus an indication that economic conditions may have an impact on cost to income ratios with lagged effect. 23 Further research appears necessary before it will be appropriate to draw conclusions Another problem with using cost to income ratios as a driver for bank management is that the ratio for any particular bank is not necessarily stable. This is illustrated in analysis of figures by McCoy, Frieder and Hedges [(1994), pp ], who argue that a bank s position on the cost to income ratio spectrum in 1988 was not a basis for predicting its position on the spectrum in They suggest that cost to income ratio will tend to revert towards some standard industry-wide mean value. There are, however, some banks with low cost to income ratios in both time periods, such as Fifth Third Bancorp and Wells Fargo. These banks have established reputations for low cost was a year of negative growth, and by 1992 the economy had started to recover. 19

20 to income ratios, which may reflect factors peculiar to the markets in which they operate. 24 It can be argued as follows that a cost to income ratio advantage should not normally be sustainable. Suppose a bank had a lower cost to income ratio because of higher levels of income. Assuming that banking markets are contestable, this income advantage should be eliminated over time by the entry of new competitors [Oliver (1996)]. Alternatively, if a cost to income ratio advantage arose from lower costs, the bank would be expected to undercut its competitors and grow its book, so as to increase overall profits. Its income advantage would thus decrease while its costs increased, driving its cost to income ratio back towards some market norm or equilibrium level. A logical consequence of this process is that, other things being equal, the more competitive a banking market, the higher will be cost to income ratios. 25 In a more competitive market, revenues will be lower, and banks will therefore have to strive to achieve ever lower levels of cost. It can thus be argued that any sustainable advantage in cost to income ratio must be a result of continuing successful innovation to reduce costs or maintain higher income. Conversely, successful innovative banks amongst the United States regionals and super-regionals do not necessarily have low cost to income ratios. One successful bank with a relatively high cost to income ratio is Norwest, where the focus has been on getting the business right first, and then letting increased business volumes overcome the apparently high cost levels In the case of Wells Fargo, this cost to income ratio advantage was significantly undermined by the subsequent acquisition of First Interstate. This is consistent with the argument below where a bank with a cost to income ratio advantage uses that advantage to expand its business. 25 There is therefore a question as to why the relatively low cost to income ratios recorded for Canadian and Swiss banks (in Table 1) persist. 26 Cost to income ratio has reduced in stages from 74.0% in 1993 to 62.8% in 1997, which is still notably higher than the ratio suggested by James et al (1997), but Norwest has still been in a position to acquire Wells Fargo. 20

21 The preceding discussion makes us ask why the National Australia Bank has the lowest cost to income ratio amongst the Australian majors. To some extent, the low ratio is a result of a focus on cost reduction, but it also reflects the range of different markets in which it operates, and the nature of its business in those markets. 27 By comparison with the ANZ, for example, it must be assumed that National Australia s retail operations in the United Kingdom and Ireland are lower cost than the ANZ s essentially wholesale and trade offices in Asia and retail banking in India. There must also be a question as to whether the advantage enjoyed by the National Australia Bank group can be sustained, particularly in view of the marginal increase in cost to income ratio in recent years. The danger with emphasis on cost to income ratios is that bank managements are inclined to assume that the response to any downturn in bank profitability is to reduce costs. This will often involve essentially knee-jerk short-term reactions, rather than reflection on the bank s business 28 or re-engineering to achieve sustainable long-term cost reduction. A classic example of this approach has been action by banks to make older, more-experienced managers redundant, which loses a body of corporate knowledge from banks. This means that in due course the banks have to re-learn lessons from the past which the redundant managers would have known. Another example is the tendency to cut training budgets in the face of so-called cost crises, which may reflect reductions in income, nothing to do with banks cost structures at all. An example of a call for cost-cutting in response to a reduction in profitability was in the National Bank of New Zealand after poor results for the first half of These poor results largely reflected problems with the bank s interest margin, although the bank s managing director, Sir John Anderson, also identified rising costs as an issue. 29 The proposed solutions 27 The National Australia Bank is also a practitioner of benchmarking in respect to cost to income ratios, with a policy of seeking to identify good practice in individual banks and then spread these throughout its portfolio of banks. For a discussion of benchmarking, see McCoy, Frieder and Hedges (1994), pp The bank s business may legitimately be higher cost than some of its peers, and cost elimination may undermine other business objectives. Refer Asher (1994), McLean (1994). 29 As reported in the New Zealand Herald, 19 August 1996, p 1 of Section 3, in a story 21

22 were two programmes aimed at reducing cost levels, one of which, called value management focused on re-engineering to take account of changes in markets. The other programme was described as cost containment, and involved staff reductions over the next six months, with redundancies as necessary, although losses by attrition were to be preferred. It is arguable that the National Bank s focus should have been on trying to restore its income, at least to the levels achieved by its competitors. The focus on costs is taken even further by James et al (1997), when they argue that banks must reduce their cost to income ratios below the 55 to 60% level if they are to avoid being taken over. This is because higher cost competitors are relatively disadvantaged as price competition drives down margins. Against that background it is interesting to review the actual trends among the 100 largest banks in the United States over the period 1986 to Over this period bank incomes relative to assets have increased (contrary to Australasian experience), as have costs relative to assets. There has been a small reduction in cost to income ratios (from 66% to 64%), as incomes have risen faster than costs. This discrepancy between the actual figures and the assumptions underlying the article by James et al (1997) casts some doubt on the validity of the article s conclusions. SUMMARY The stress on cost to income ratio in banking has both strengths and weaknesses. It is not bad for banks to focus on their costs, and, other things being equal (which they very rarely are), banks with lower cost to income ratio are likely to be more profitable. It is also more realistic National Bank feels squeeze on profitability. In fact, according to the author s analysis, operating expenses to assets decreased significantly (although the bank s cost to income ratio increased). The major impact was on income, with net interest income decreasing.46% (from 2.49% to 2.03%) and non-interest income decreasing.46% (from 1.72% to 1.26%, largely on the back of reduced profits from sale of securities and losses from securities trading). 30 Analysis is based on OECD (1997) and Nelson & Owen (1997). 22

23 to compare cost to income ratios across international boundaries than it is to compare cost to assets ratios, which are even more sensitive to structural differences between banks. There are, however, grave risks in comparing cost to income ratios internationally, as they are inclined to be distorted by differences in the structure of banks operations. 31 There are other warnings in respect of cost to income ratios. If banks calculate them themselves, the calculations may not be transparent. If one is using cost to income ratios to compare banks, one needs to acknowledge that there are a number of factors which may cause them to vary from one bank to another. Also, adverse movements in cost to income ratios are often used by bank managements as excuses for short-term cost cutting, without regard to the cause of the changes. Cost to income ratios are important to commentators on banking, however, and bank managements are often forced to focus on them and comment on them. They cannot be ignored, and it is therefore as well that they are understood. It is hoped that this paper will make some contribution towards spreading that understanding. Bibliography: Asher, J. (1994, September). Can efficiency go too far. ABA Banking Journal. Pp 43-44, 46, 48. Bekier, M. M. & Nickless, S. (1998, 1). Banks need fewer checks, not fewer branches. The McKinsey Quarterly. Pp Buttle, J. (1998, 17 August). Delivery channels, earnings spread to determine winners. Australian Banking & Finance. Pp Davis, S. (1994, February). How banks control costs. Banking World. Pp Greising, D.; Galuszka, P.; Morris, K.; Osterland, A. & Smith, G. (1998, 27 April). $1,000,000,000,000 banks. Business Week. Pp 32-36, Harris, P. (1996). The Changes Banks Face. Wellington: Finsec. 31 As Toevs and Zizka (1994) comment, even if attempts are made to normalise cost to income ratios for differences in capital structure or other factors, there will still be other structural differences which cause cost to income ratios to differ. 23

24 James, M.; Mendonca, L T.; Peters, J. & Wilson, W. (1997, 4). Playing to the endgame in financial services. The McKinsey Quarterly. Pp Madewell, G. (1997, April). Best stick to cash. The Banker. Pp 24, 26. McCoy, J. B.; Frieder, L. A. & Hedges, R. B. Jr. (1994). Bottomline Banking. Chicago, Illinois: Probus Publishing. McLean, J. (1994, May). Look after the customer. The Banker. Pp Neagle, T. F. (1995, March). Automated underwriting Down Under. Mortgage Banking. Pp Nelson, W. R. & Owen, A. L. (1997, June). Profits and balance sheet developments at U.S. commercial banks. Federal Reserve Bulletin. Pp OECD (1997). Bank Profitability. Development. Paris: Organisation for Economic Co-operation and Oliver, R. (1996, December). The contestability of the New Zealand market. New Zealand Banker. Pp Osborne, J. (1995, Summer). A case of mistaken identity: the use of expense/revenue ratios to measure bank efficiency. Journal of Applied Corporate Finance. Pp Stock, J. (1998, March). ANZ Bank - a case study. Equity (an ASA Publication). Pp Toevs, A. & Zizka, R. (1994, Summer). Straight talk on bank efficiency. Journal of Retail Banking. Pp Tripe, D. (1996, September). The cost to income ratio. New Zealand Banker. Pp 7-8. Wall, L. D. (1983, September). Why are some banks more profitable? Federal Reserve Bank of Atlanta Economic Review. Pp Welch, P. (1994, December). Counting the costs. Banking World. Pp