Abstract
Individual transferrable quotas (ITQs) are used in Norway’s fisheries, but only partly and with restrictions on transferability. This paper examines how this has affected return on capital. The return on boat capital in three of four fisheries examined has increased since ITQs were introduced. A part of this increase is due to an increase in value of fish landings, but accounting for that still leaves room for increases due to ITQs. Return on total capital (including quota value) has either stagnated or fallen. This is consistent with the rate of return in ITQ fisheries being comparable with other industries in the long run. (JEL Q22, Q28)
I. INTRODUCTION
Many fisheries the world over have now been put under some form of individual transferable quota (ITQ) management.1 In Norway this has happened gradually and with much reluctance. Quotas of boats that are decommissioned can be transferred to other boats, but with restrictions. Quotas and other fishing rights such as concessions that have been transferred appear in the accounts of firms as assets whose value is determined by the price at which they were purchased. Purchased quotas can be retained only for a limited number of years, usually 20, and can be depreciated over that time.
The scenario in fisheries that have been put under ITQ management is by now well known. There is consolidation of the fishing fleet as boats are decommissioned and their quotas accumulate on fewer boats. Boats that remain in the industry increase their revenue and obtain some resource rent. But quotas are purchased at a price. When these outlays are included in the capital base of boats, the return on their total capital declines. If quota trade is unrestricted, return on capital in a quota-managed fishing industry will in the long run be the same as in other industries, with an appropriate allowance for risk. Hence, the gains for the industry will be transient, with the resource rent becoming capitalized in quota values. Copes (1985), in his critique of ITQs, described this as a “transitional gains trap.”
How has this played out in Norway’s fisheries?2 Has the return on capital invested in boats improved over time as the transfer of quotas and other fishing rights has been liberalized? Has the return on total capital, that is, also including quota purchases, nevertheless remained the same, as predicted by the transitional gains trap? This is examined in the present paper. Care is taken to isolate the effect of changes in the value of fish catches apart from general inflation. This is important, as both fish catches and fish prices have trended upward, even if subject to substantial fluctuations.
The empirical literature on the effects of ITQs has primarily focused on changes in productivity and not on changes in the return on capital.3 An exception is a paper on the Norwegian purse seine fishery in the 1980s (Flaaten, Heen, and Salvanes 1995), a time period when quota transfers were much more circumscribed than they are today. They found that boats bought together with quotas were more expensive than the ones whose owners had gotten their quotas for free, and so the return on capital for the latter was appreciably higher. Matthíasson (2012) discusses a related issue, the asset value of quotas in the Icelandic fisheries and how it was affected by the stock market bubble.
II. THE DATA
We use data from the annual cost and earnings studies conducted by the Directorate of Fisheries.4 The data pertain to individual fishing boats, but in anonymous form. The data are from a sample of boats, which changes over time, with boats going in and out of the sample over time. The years covered by our data set are 1985–2013. The sample is meant to be representative, but we are not in a position to evaluate this independently. For some boat groups we study (purse seiners and trawlers) the sample covers almost all boats in the group. The data are reported by the fishing enterprises in question and audited by the Directorate of Fisheries.
We study four major groups of fishing boats in Norway: (1) purse seiners over 90 feet; (2) trawlers fishing mainly for cod, haddock, and saithe; (3) boats using conventional gear (long lines, hand lines, or nets) and fishing offshore (boats over 28 meters); and (4) coastal boats using conventional gear (boats less than 28 meters). This covers most of Norway’s fisheries.
The data on earnings and operating costs are straightforward. Income of capital is defined as earnings from fishing less operating costs. Operating costs comprise labor costs, social security fees and other taxes, and fuel and other intermediate goods. We include maintenance costs for the boat and fishing gear in the operating costs, but not depreciation, neither of the boat itself nor its fishing rights. Depreciation costs depend on taxation rules and do not necessarily reflect real depreciation. Fishing rights do not depreciate in any real sense, but fish quotas that have been purchased can be retained only for a limited time, usually 20 years, which is the reason for their depreciation for tax purposes. Hence we are dealing with gross capital income, from which the investment cost of fishing boats and fish quotas will have to be recovered.5
III. CAPITAL VALUE OF BOATS
To calculate the rate of return on capital, we need the value of capital invested in fishing boats and other equipment as well as fishing rights. This is where we encounter the greatest problem with the data at hand. The cost and earnings studies report the book value of fishing rights only from 2002 (these values will be discussed later) and the book value of fishing boats from 1994. The book value of a boat is, however, not necessarily an adequate expression of its real value. Boat values are depreciated over 13 to 18 years, but that depreciation may have little to do with real depreciation of the boats; there are many boats with zero asset value in the accounts, and many are several decades old. Hence, even when we have the book value of a boat, it may tell us very little about what it is worth.
A way to get around this is to use the insurance premium to estimate the real value of boats. According to sources in the insurance companies the insurance premiums are related to the market value of boats, but with certain rebates or penalties; rebates are given for long periods in harbor and penalties applied for accidents that have occurred. Looking at insurance premiums of a given boat over time we sometimes see variations that are impossible to explain by presumptive changes in the market value of the boat over time. The insurance premiums thus undoubtedly reflect the capital value of boats imperfectly, but better proxies are unfortunately not available for all boats and years in our data set.
The information that the insurance premium is related to the market value of boats is confirmed by regressing the insurance premium on the market value of boats in cases where we have information on the value of boats when new. For boats that are only a few years old and for which we have depreciation records, we can reconstruct the book value when new. Regressing the insurance premium on the boat value when new gives the results shown in Table 1. The insurance premium increases with the value of boats for all four groups of boats, but the regression coefficient is significant only for three (purse seiners, trawlers, and conventional coastal boats), while it is insignificant at any reasonable level for conventional offshore boats. Only a small part of the variations in insurance fees is explained by differences in value of boats, except for conventional coastal boats, but this group is more numerous and varied in terms of value when new than the other three types.6
Regression (y = a + bx) of Insurance Premium (y) on Book Value of Boats When New (x)
Using the value of boats when new to explain the insurance premium implies that the value of boats as capital assets remains constant over their age, which is not necessarily true, but whether it increases or decreases in nominal terms is not obvious. Wear and tear, as well as technological progress, reduces the real value of boats, but because of general inflation the nominal value of boats could well increase even if their real value falls, provided it rises at a lower rate than the rate of inflation. On the other hand, boats could be so well maintained that their real value is preserved; many boats in the fleet are decades old, but most of the contents of the hull (engines and other equipment) has probably been replaced at least once. Annual maintenance costs are substantial: 2% to 4% of original boat value adjusted for inflation.
To investigate this, we regressed the logarithm of the insurance premium on boat age, both for nominal and deflated insurance premiums (using the consumer price index as a deflator).7 If insurance premiums rise with age, it would be due either to increasing cost of insurance irrespective of the value of boats or to a rising nominal value of boats, given that the insurance premium is related to the value of the boat. If, on the other hand, insurance premiums stay constant over time, it would be an indication that the value of boats also remains constant over time, and if the insurance premiums fall, the value of boats would also fall. As Table 2 shows, we get mixed results for the four boat groups. For purse seine and conventional coastal boats we find that the insurance premium rises with the age of the boat by 1% to 2% per year. This indicates that the boats are sufficiently well maintained for their nominal value to increase over time. Regressing the deflated insurance premium on boat age produces insignificant coefficients, so we conclude that the real value of the boats is preserved over time, with the nominal value rising at approximately the rate of inflation. The fact that we get different results for the four boat groups indicates that a general rise in insurance costs is not the reason for rising insurance premiums with age.
Regression (ln y = boat dummies + at)of Logarithm of Insurance Premium (y), Nominal and Deflated, on Age of Boats (t) and Dummy Variables (Not Reported) for Boats for Which We Have Value When New
For the other two groups, the trawlers and the conventional offshore boats, we find no significant increase in insurance premiums with age, but a significant decline if the insurance premiums are deflated by the consumer price index. This indicates that the real value of boats in these groups erodes at about the rate of inflation, such that their nominal value is preserved over time
IV. RENEWAL OF BOATS
We have found that the real value of some boats seems well enough preserved for their real value to stay constant over time. Does this mean that boats will be maintained forever and never replaced? Not at all, even if they may be maintained for a long time. The problem at hand is one of optimal replacement of plant and machinery, reviewed and further analyzed by Dobbs (2004). There is evidence that maintenance costs rise with the age of boats. Regressing the logarithm of maintenance costs divided by boat value adjusted for inflation on age yields a positive coefficient for all boat groups (Table 3).8 If maintenance costs rise with the age of the boat in a geometric fashion at a rate k, the maintenance cost at time t (ct) would be
Regression (ln(y1/y2)=a + bt) of Logarithm of Maintenance Costs (y1) Divided by Inflation-Adjusted Value of Boats When New (y2) on Age (t)
For an optimal renewal program for the boats, the present value (V) of new boat purchase (X) and all maintenance costs would be
where r is the discount rate and T is the boat’s age at replacement.9 Finding the optimal age for boat renewal involves minimizing V with respect to T. Taking the derivative of V with respect to T and setting equal to zero gives
and implicitly, the solution for optimal T. Writing this as
provides a straightforward interpretation. Buying a new boat at time T and following the optimal replacement program generates a cost equal to the present value of boat purchases plus maintenance cost for each boat, V. Perturbing the replacement schedule and postponing replacement generates a cost equal to the present value of maintenance costs at age T, which is c0ekT/r. Hence, boats will be replaced when their maintenance costs outweigh the costs of a new boat.
Figure 1 shows how the optimal age of replacement depends on the discount rate. Raising the discount rate from 1% to 10% increases the optimal replacement age from about 30 years to about 40 years for most vessel groups. There is hardly any difference between the optimal replacement ages for three of the vessel groups (purse seiners, trawlers, and conventional coastal boats), but for the conventional offshore boats the replacement age is appreciably lower. As we have seen, the real value of the trawlers does not seem to be maintained over time, and if we relate the maintenance costs to their nominal value we get the curve labeled “Trawlers 2” in Figure 1. This shortens the optimal replacement age by 7 to 12 years, depending on the discount rate. This is as expected; if good maintenance is not fully able to maintain the real value of the boats, they will be replaced earlier.
How the Age of Replacement (T) Depends on the Interest Rate (r)
The optimal replacement age shown in Figure 1 roughly agrees with the observed age of boats, except for conventional coastal boats. We do find a few boats older than 40 years, but these are exceptions (for average age of all boats in the sample and its standard deviation, see Appendix Table A1). Among the conventional coastal boats we find vessels that are more than 60 years old. This is a very heterogeneous group, perhaps too much so to be aggregated in a meaningful way.
V. RETURN ON BOAT CAPITAL
As Table 1 shows, the insurance premiums are correlated with the value of boats when new, but using the results in Table 1 nevertheless leads to unrealistic estimates of boat values in some years. The reason is that the insurance premium can vary a great deal from one year to another in ways that cannot be explained by changes in the market value of the boat, for reasons that have already been mentioned and on which we have no information. We therefore estimated the following equation without a constant:
where Ki is the value of boat i when new and Ii is the insurance premium (Table 4). We then used a to estimate the capital value of each boat for each year. Because the year-to-year variations in the insurance premiums are in some cases very substantial and not explainable by changes in the market values of the boats, we use the average value of K predicted for the years we have observations for. For trawlers and conventional offshore boats we use nominal values and keep the boat value unchanged over time. For the other two, purse seiners and conventional coastal boats, we use deflated insurance premiums to estimate the initial value of the boat and then appreciate it over time by the consumer price index. The justification for this is the finding that the insurance premiums for these boat groups (purse seiners and conventional coastal boats) increase with age by approximately the rate of inflation, but premiums are uncorrelated with age for the other two groups.
Coefficient of Proportionality between Insurance Premiums and Value of Boats When New
Even if the value of boats when new estimated from the insurance premiums can vary a great deal from the purchase value of boats for individual boats, the average return on boat capital emerging from the two sets of boat values is for the most part fairly similar. This is illustrated in Figure 2. The biggest discrepancy occurs for conventional offshore boats in the early 1990s, but the number of boats for which we have book value when new is very small: 1 to 3 boats each year.
Average Return on Capital for Boat Values Estimated by Book Value as New and Insurance Premiums
Figure 3 shows the return on capital invested in boats for the four boat groups. For three of them (purse seiners, trawlers, and conventional offshore boats) there has been a substantial increase since the mid-1980s. Over the same period, quota trading rules have been liberalized and the number of boats has shrunk. It is not entirely straightforward, however, to relate the improvement in return on capital to the relaxation of these rules. For trawlers and purse seiners, the main relaxation came in 1996 when boat owners were allowed to purchase quotas from decommissioned boats.10 The improvement in the rate of return on capital in the purse seine fleet began at about this time, but for the trawlers this did not happen until after 2003 (quota transfer rules were further liberalized for this group in 2000). For the conventional offshore boats the quota transfers were liberalized in 2000, but the improvement in return on capital began a few years later, from 2004 onward. For these boats the improvement occurred rapidly; it seems to have run its course already in 2006. The development of the conventional coastal fleet does not conform to the liberalization of quota transfer rules; the main improvement occurred from the late 1980s to the turn of the century, but the quota transfers were permitted only from 2004 for boats 15 to 28 meters and later for smaller boats. Nevertheless, the stagnation in returns on boat capital continued after 2004.
Return on Boat Capital for Four Boat Groups, 1985–2013: “Adjusted” Shows Rate of Return Divided by the Index of Catch Value Deflated by the Consumer Price Index
The second curve shown in these diagrams, the one labeled “adjusted,” shows the rate of return on boat capital divided by an index of catch value. This index is constructed from the value of annual fish catches of the most important species11 deflated by the consumer price index. Hence, an increase in this index shows an increase in catch value over and above what would follow general inflation. The shape of the adjusted curves is rather similar to the unadjusted one, but we see that a large part of the improvement in the rate of return on purse seiners, roughly about one-half, is due to larger catches and higher fish prices, rather than other factors such as liberalized quota transfer rules. Still, a substantial increase in return on capital must be ascribed to factors other than increase in catch value, the most obvious of which is the liberalization of quota trading rules, which has permitted a concentration of fish catches on fewer boats. For the conventional coastal boats we see no improvement in profitability since the quota trading rules were liberalized in 2004.
It could be argued that “deflating” the return on capital by an index for fish catch values (after adjusting for inflation) is not an appropriate method to correct for changes in catch value. For given costs, an increase in catch value will disproportionally increase the revenues after operating costs and hence increase the return on capital more than proportionally. But operating costs are not independent of the catch value, because labor is not paid a fixed wage but rather a share of the catch value. Since labor costs are the most important part of operating costs, it is not farfetched to argue that the return on capital will change roughly in tandem with the catch value.
VI. RETURN ON TOTAL CAPITAL
Boats enjoying exclusive fish quotas and other fishing rights will obtain better results than other boats. When such rights can be bought and sold, boat owners will invest in such rights up to the point of obtaining a return on their investment equal to what they could obtain otherwise, eventually adjusted for risk. We would, therefore, not expect to see any long-term increase in the return on total capital invested as a result of more liberal quota trading rules, just normal return comparable to other industries, eventually adjusted for risk. Figure 4 shows what has happened to the average rate of return in the four fisheries we have been observing. For all four boat groups a gap has developed between the rate of return on boat capital only (that is, capital income divided by the value of boats) and the rate of return on capital invested both in boats and in fishing rights. The rate of return on total capital12 has in fact fallen for the purse seiners and conventional coastal boats since 2001. For the other two, trawlers and conventional offshore boats, the gap between the two rates has widened while the rate of return on total capital has increased or stagnated.
Return on Capital Invested in Boats Only and on Capital Invested in Boats and Fishing Rights
Looking beyond averages, we may note, first, that there is large variability in the rate of return across boats; in any single year the confidence intervals overlap (not shown in the figures). Nevertheless, it is unwarranted to conclude that the gap between the two rates of return is nonexistent and not widening. A regression of the difference between the two rates of return on time for all boats shows that this difference has been rising significantly over time (Table 5). This conclusion is further supported by regressions for each boat individually (Table 6; only boats with four or more observations are considered). For half of the boats or more we get a significantly rising difference over time, except for conventional coastal boats, for which only a quarter of the boats show a significantly rising difference over time. In this last case this is partly due to the fact that we have many fewer observations for the individual boats than for the other three boat groups.13
Value of the Coefficient b in the Regression y = a + bt Where y Is the Difference between Return on Boat Capital Only and Total Capital and t Is Calendar Year Minus 2001
Results from Regression as Reported in Table 5 Run for Individual Boats with Four or More Observations; Number of Boats with Significant versus Insignificant Regression Coefficient (b)
As Figure 4 shows, the two rates of return were equal in 2001. This is simply because fishing rights were not registered as separate asset category before 2002. But, as is also evident, the value of fishing rights was quite low in the first years after 2001, partly because fewer rights had been traded and partly because they traded at a lower price in these early years. We have taken the value of fishing rights from the accounts of the fishing boats, as reported in the cost and earnings studies. Fish quotas can be depreciated over the lifetime of the quotas, so for these rights we use the reported value each year plus depreciation of the quotas in earlier years; this accounts for the money originally expended on these assets. Because all boats are not represented in all years, we may have missed one or a few years of depreciation and so undervalued the quota rights. This will certainly have happened for quotas purchased in 2001 and earlier.
VII. CONCLUSION
There seems little doubt that the quota trade in Norway’s fisheries, albeit limited, has improved the utilization of capital invested in fishing boats. This is, needless to say, a substantial economic gain, from the point of view of the national economy. The price paid for fish quotas does not represent a real cost but a transfer from one party (the buyer of quota) to another (the seller of quota); the amount of fish to be taken remains the same. The only case in which such transaction would represent a real cost is if the money could otherwise have been invested in a more profitable opportunity. The fact that the quota buyer chose not to do so indicates the absence of any such opportunity. For the boat owners the gains are less evident. They have spent substantial sums on purchasing fish quotas so that the overall return on their total invested capital has changed much less than the return on boat capital only, but it still seems quite comfortable; in recent years it has been 5% to 6% for boats fishing with conventional gear, and 10% to 15% for trawlers and purse seiners (as stated earlier, these are gross rates, out of which investment costs must be recovered). A large share of the improvement in return on boat capital is due to the higher value of fish catches, but nevertheless, the contribution of fish quota trades is substantial. One group that for certain has gained is the boat owners who have left the industry and sold quotas they originally got for free.
Acknowledgments
I am grateful to the Directorate of Fisheries for having provided data for individual boats in anonymous form and to Tove Aasheim, Directorate of Fisheries, for answering queries.
APPENDIX
Average and Standard Deviation of Variables Used in the Statistical Analysis
Footnotes
The author is professor emeritus, Norwegian School of Economics, Bergen, Norway.
↵1 A recent review is provided by Arnason (2012). There is a concise, but by now dated, review by Grafton (1996). Classic references on fisheries economics include work by Gordon (1954) and Clark (1976). On ITQs and investment in fishing boats, see Hannesson (2000).
↵2 In an earlier study, this author looked at the effects of ITQs in the Norwegian purse seine fishery (Hannesson 2013), but that study ignored the effect of the larger fish quotas and higher fish prices that this article deals with explicitly.
↵3 The empirical studies of improvements in productivity after the introduction of ITQs are too numerous to mention here in full. Examples include work by Grafton, Squires, and Fox (2000) and Squires et al. (2010). At least one study (Walden et al. 2012) concludes that productivity gains from ITQs have not been sustained in one particular fishery, but apparently not because of the quota management as such. Costello et al. (2012) conclude that fish quotas, not necessarily transferable, can save fisheries from collapsing.
↵4 Aggregate data are available at www.fiskeridir.no/Yrkesfiske/Statistikk-yrkesfiske/Statistiske-publikasjoner/Loennsomhetsundersoekelse-for-fiskefartoey.
↵5 The data at hand report values for individual fishing boats. The question has been raised as to how reports from companies owning more than one boat are dealt with (allocation of insurance costs, for example). Most fishing enterprises in Norway own only one boat, due to the requirement that boats should be owner-operated. The Directorate of Fisheries audits the data for appropriate allocation of fixed costs and insurance costs, among other things.
↵6 The intercept in the regression is fairly similar for three of the boat groups (trawlers, purse seiners, and conventional offshore boats), whereas for the coastal boats we get an intercept close to zero, albeit significantly positive. For the latter group we have some observations of boat values and insurance premiums close to zero (minima of 0.2 and 0.002, respectively) while for the other three the minima are 17– 23 and 0.04–0.2. One could interpret this significant intercept as an indication that there is some fixed part to the insurance premium, with the rest varying with the boat value.
↵7 For this we used the subsample of boats for which we have information on value when new.
↵8 Also here, the subset of boats for which we have information on value when new is used.
↵9 This is equivalent to Dobbs’s expression for the deterministic case, except that the salvage value is ignored. In a perfect second-hand market there would be no point in replacing an existing boat with a second-hand one.
↵10 Before that time it had been possible in certain cases to transfer quotas permanently between boats in these groups. In the purse seine fishery, concessions could be transferred from decommissioned boats. Quotas were determined on the basis of the concession capacity, but in a regressive manner.
↵11 These species are cod, haddock, and saithe for the conventional boats and the trawlers, and herring, capelin, mackerel, and blue whiting for the purse seiners. The catch quantities have been collated from the minutiae of the regulations committee (Reguleringsutvalget) from various years from 1985 to 2013 (available at https://brage.bibsys.no/xmlui//handle/11250/117273), while the prices are simply the total value of landings divided by quantity.
↵12 We define total capital as capital invested in boats and fishing rights. In the cost and earnings studies there is a residual of assets in addition to those two that is included in total capital. Therefore, the rate of return on total capital in the cost and earnings studies is less than here. There are reasons for not including this residual in the concept of capital necessary for operations; this category may include reserve funds or investment in financial assets for tax purposes. In any event, this category can vary substantially from year to year in ways that appear unrelated to the fishing operations.
↵13 Since the difference between the two rates of return can be in only one direction, we should apply a one-sided test, so the critical 5% level we are using really is 2.5%.
