P. J. Romilly
Department of Accountancy and Economics Dundee Institute of Technology
Department of Management University of St. Andrews
In recent years there has been a gradual shift away from the traditional lecture approach (passive learning) towards a more student-centred approach (active learning) in higher education. There is also a recognition that theory should be integrated with applications. Judge (1990) provides a wide range of spreadsheet applications aimed mostly at the introductory (first year) undergraduate level. This paper aims to show how an integrated approach to teaching and learning using spreadsheets can be extended to intermediate macroeconomics. The approach is illustrated using a spreadsheet version of the Mundell-Fleming open economy IS-LM-BP Model.
The aims of this paper are twofold: firstly, to explain the advantages of using spreadsheets in teaching macroeconomics; and secondly, to demonstrate the construction and applications of a spreadsheet macroeconomic model which can be used at intermediate level in the economics components of specialist (economics) and non-specialist (e.g. business studies or accounting) degree courses. The model which is used in this paper is the fixed price, open economy IS-LM-BP model; a full exposition of the model can be found in textbooks such as Greenaway and Shaw (1988), Mankiw (1992) and Dornbusch and Fischer (1993).
The paper is structured as follows. Section 1 outlines the advantages of using spreadsheets for teaching macroeconomics. Section 2 describes how users can set up an operational version of the IS-LM-BP model for themselves using a spreadsheet. Finally, section 3 provides some illustrations of ways in which the model can be applied.
Advantages of using spreadsheets in teaching macroeconomics
Most textbook expositions of macroeconomics rely heavily on comparative statics analysis of the equilibrium properties of a sequence of stylised macroeconomic models, progressing from the income expenditure model, through closed and open economy versions of the IS-LM model, to various possible formulations of the aggregate demand-aggregate supply (AD-AS) model. However,this approach can leave unanswered a number of important questions about the dynamics of macroeconomic models. By concentrating on the static equilibrium properties of models, little attention is focused on how the models adjust when an equilibrium position is disturbed, what time paths are followed by individual variables, and how long it is likely to take for a new equilibrium position to be achieved.
The approach advocated in this paper allows students to begin to explore possible answers to these types of question by constructing a dynamic counterpart to a static textbook model for pedagogic purposes.Although the parameter values and dynamics are not derived econometrically, their specification is such that the model can be used to provide insights into actual events such as the contractionary monetary policy of the Medium Term Financial strategy and the UK's withdrawal from the Exchange Rate Mechanism (see section 3). By using a spreadsheet to build such a model, students can examine for themselves, step by step, the implications of changes in policy variables and other exogenous parameters on all other variables in the model, and gain a clearer understanding of the processes driving the economy at the macroeconomic level than can normally be obtained from a static model. To summarise, in our experience this approach offers the following advantages:
(a) Users gain a better understanding of the assumptions and internal relationships between the various parts of the model than is normally obtained from studying diagrams showing shifts (for example) in IS and LM curves.
(b) By presenting the equations in a form which permits users to build the model for themselves, the assumptions underlying the model are revealed with greater clarity than is often the case with pre-programmed learning packages.
(c) Users can run simulations of the effects of changes in the exogenous variables, and investigate the implications of such changes for all other variables. Issues concerning model stability and patterns of adjustment can be addressed and explored in an interactive manner.
(d) Users can enhance their spreadsheet skills while studying macroeconomics.
(e) These objectives can be achieved with very little use of mathematics, making the method appropriate for both specialist and non-specialist students of economics.
A possible disadvantage is that some laboratory time (and supervision) is required if students are to program the model for themselves; however, these costs could be avoided if the instructor is willing to prepare in advance and then circulate a 'master' copy. While we do not suggest that laboratory sessions using spreadsheet models of the type described here can ever replace conventional lectures and tutorials, in our experience they can provide an extremely useful complement by helping to explain the properties of the textbook models in an interesting and vivid manner.
Programming the spreadsheet version of the IS-LM-BP model
The algebraic specification of the spreadsheet version of the IS-LM-BP model is as follows:
(1) Consumption C(t) = a + b[Y(t) - T(t)]
(2) Planned investment I(t) = d - f i(t)
(3) Government expenditure G(t) = G
(4) Taxation T(t) = v Y(t)
(5) Exports X(t) = x/e(t)
(6) Imports Z(t) = z e(t) Y(t)
(7) Aggregate expenditure AE(t) = C(t) + I(t) + G(t) + X(t) - Z(t)
(8) Current account CA(t) = X(t) - Z(t)
(9) Capital account KA(t) = k[i(t) - n]
(10) Demand for money L(t) = g + h Y(t) - j i(t)
(11) Supply of money M(t) = M(t-1) + CA(t-1) + KA(t-1)
(12) Income/output Y(t) = AE(t-1)
(13) Rate of interest i(t) = i(t-1) + s[L(t-1) - M(t)]
(14a) Fixed exchange rate e(t) = e
(14b) Flexible exchange rate e(t) = KA + û[KA(t)*KA(t) + 4xzY(t)]/[2zY(t)]
Equations (1) to (6) embody a set of standard 'Keynesian' assumptions concerning the determinants of the components of aggregate demand; (7) and (8) are self-explanatory identities;(9) specifies the capital account in terms of the differential between the domestic and world interest rates; (10) specifies the demand for money; and (11) to (14) are adjustment equations for the domestic money supply, real income, the interest rate and the exchange rate respectively. In (11) the domestic money supply responds directly to balance of payments surpluses or deficits which may arise if the exchange rate is fixed. (12) embodies an assumption that aggregate supply responds directly to changes in the level of aggregate demand, subject only to a time lag of one period. (13) causes the interest rate to adjust to imbalances between the demand for and supply of money (with a lag of one period on the demand side). Finally, (14a/b) determine whether the exchange rate is fixed (14a) or flexible (14b); in the case of the latter, the exchange rate adjusts to eliminate any balance of payments surpluses or deficits instantaneously (i.e. before they actually materialise).
Tables 1 and 2 illustrate the layout of the model after it has been entered into a spreadsheet. We recommend that the model is set up across two screens. Table 1 shows the first screen. This contains the numerical values of all the policy variables and other exogenous parameters both before (at time t=0) and after (from time t=1 onwards) the occurrence of a 'shock' which is to be investigated. The parameters and policy variables at t=0 (column B) are treated as constant, and cannot be changed; those from t=1 onwards (column C) can be changed to any values which the user wishes to feed into the model. Table 2 contains the numerical values of the macroeconomic variables themselves (as defined in equations (1) to (14)) in each time period from t=0 onwards. For illustrative purposes, Tables 1 and 2 show the effects of an increase in government expenditure (from G=250 at t=0 to G=300 from t=1 onwards). A full explanation of how to set up these tables is available from the authors on request. Only a very elementary knowledge of spreadsheet programming is required.
[Tables 1 and 2 here]
We illustrate the behaviour of the model by investigating the impact of an increase in government expenditure under a fixed exchange rate regime, as shown in Tables 1 and 2. The static equilibrium analysis of such a change would normally proceed along the following lines.
Initially, the increase in government expenditure shifts the IS function from IS to IS in Figure 1, and the economy moves from its original equilibrium at A to a new, intermediate equilibrium at the intersection of IS and LM at B. This is a position of balance of payments surplus, since B lies above the BP function. Therefore, central bank intervention in the foreign exchange markets in the form of sales of sterling is required to prevent the exchange rate from increasing. This leads to an increase in the domestic money supply, which causes the LM function to shift from LM to LM . The economy moves from its intermediate position at B towards a final equilibrium at the intersection of IS , LM and BP at C.
[Figure 1 here]The Effects of an Increase in G from 250 to 300
Table 2 shows the details of the process by which the dynamic model adjusts from A to C following the injection of government expenditure.The initial increase in aggregate expenditure at t=1 leads to an increase in income (Y) at t=2, which stimulates expenditure on imports, opening up a current account (CA) deficit at t=2. The increase in Y also causes an increase in the demand for money (L) and an increase in the interest rate (i). By t=4, the capital account (KA) improvement caused by the increase in i begins to exceed the CA deficit, so the money supply (M) starts to increase, exceeding its initial value of 500 for the first time at t=6. This monetary expansion then causes i to begin to fall back from t=7 onwards, which has further repercussions for future values of i and KA. All of the endogenous variables then pass through several cycles of diminishing amplitude, before the economy eventually settles at a final equilibrium which corresponds to C in Figure 1.
The dynamics of the adjustment process can conveniently be represented in Figure 1 by means of the 'spiral' between A and C, which is constructed by plotting each successive pair of values of i and Y (obtained from Table 2) for each time period from t=0 onwards. Although the description of the adjustment process turns out to be more complex than that provided by the 'static' analysis of the shifts from A to B to C, it also seems considerably more plausible since it provides an explicit account of the way in which the interrelationships between the individual variables drive the economy towards its new equilibrium position.
The model can also be used to explore certain macroeconomic implications of various 'real world' events, by running some fairly stylised and schematic simulations of policy changes or other exogenous shocks. Of course, the degree of realism which can be achieved is dictated largely by the scope of the model specification. In the present case, we cannot expect to simulate any effects arising from supply side influences or price flexibility, since IS-LM-BP is a demand driven, fixed price model (although a strength of our approach is that the model specification could easily be extended to include a set of supply side and price equations if required). However, the present specification can be used to simulate the demand side implications of policy decisions or other shocks, which we now illustrate using two examples.
The Medium Term Financial Strategy
Mrs Thatcher's first Conservative government assumed office in 1979 committed to a counter-inflationary strategy centred on firm control of the rate of growth of the money supply: an approach which was most clearlyarticulated in the Medium Term Financial Strategy (MTFS), initially published in 1980. Although the MTFS focussed on the rate of monetary growth, in Figure 2 and Table 3, we show a highly simplified and stylised simulation of its demand side effects, by considering a once-and-for-all reduction in the level of the money supply, from M=500 at t=0 to M=450 from t=1 onwards, using the flexible exchange rate version of the IS-LM-BP model.
In Figure 2, the LM function shifts from LM to LM and the economy moves towards an intermediate equilibrium at B, at which the balance of payments is tending towards surplus (due mainly to the capital account effects of the increase in interest rates). Upward pressure on the (flexible) exchange rate reduces the competitiveness of domestically produced goods, sending the current account into deficit and reducing the level of aggregate expenditure. The IS and BP functions shift accordingly to IS and BP , and equilibrium is restored at C, with a higher interest rate and exchange rate, and a lower level of national income. As before, the dynamics of the adjustment process are shown in detail in Table 3 (which, for reasons of space, is truncated after the first few time periods), and in summarised form by the 'spiralling' adjustment path shown in Figure 2.
[Figure 2 and Table 3 here] The Effects of a Decrease in M from 500 to 450
As Table 4 shows, a high interest rate, a strong pound and negative growth were all clear features of UK economic performance at the start of the 1980s. There has been some debate as to whether the upward pressure on the exchange rate was caused primarily by tight monetary policy (as suggested above); or by the current account effects of the UK's (more-or-less simultaneous) emergence as a net oil exporter thanks to North Sea oil; or (perhaps mostlikely) by a combination of both. The effects of reduced dependence on imported oil could easily be 'superimposed' onto the simulation shown in Table 3; for example, a reduction in the parameter z from z=0.3 at t=0 to z=0.25 from t=1 onwards causes a further appreciation in the exchange rate, over and above the levels recorded in the Table.
[Table 4 here]UK Macroeconomic Indicators 1997-86
Table 4Finally, we should comment that the possible counter-inflationary benefits of tight monetary policy are not revealed in our simulations, for the reasons discussed previously. Although it is true that the UK did eventually reap the rewards of a substantially reduced inflation rate in the mid-1980s, the extent to which this is directly attributable to the tight monetary regime of the period 1979-81 also remains a topic of considerable controversy.
____i_ e_ Dy_ ã
1977 6.4 101.2 2.6 15.9
1978 11.9 101.0 2.9 8.2
1979 16.5 107.0 2.8 13.5
1980 13.6 117.7 -2.0 18.0
1981 15.4 119.0 -1.2 11.9
1982 10.0 113.7 1.7 8.6
1983 9.0 105.3 3.7 4.5
1984 9.3 100.6 2.0 5.0
1985 11.5 100.0 4.0 6.0
1986 10.9 91.5 3.8 3.4
i = Treasury Bill yield (%)
e = Sterling exchange rate index, 1985=100
Dy = % change GDP at factor cost (constant prices)
ã = % change all items retail price index
Source: Economic Trends Annual Supplement, 1993.
Withdrawal from the Exchange Rate Mechanism
On September 16 1992, sterling was withdrawn from the European Exchange Rate Mechanism (ERM). An immediate consequence was a substantial sterling devaluation; by the end of 1992, the sterling exchange rate index was moving within a range between 12% and 15% lower than during the final days of ERM membership. This reduction was accompanied by substantial cuts in interest rates; base rates fell from 10% at the start of September 1992 to 6% in January 1993. The expansionary demand side effects are illustrated in Figure 3 and Table 5, in which we use the fixed exchange rate version ofthe IS-LM-BP model to simulate the effects of a 15% devaluation from e=1 at t=0 to e=0.85 from t=1 onwards.
In Figure 3, the current account benefits arising from improved competitiveness initially shift the IS and BP functions to IS and BP . The initial impact of the devaluation is to move the economy to the intermediate equilibrium at the intesection of IS and LM (i.e. point B, not shown). At this point there is a balance of payments surplus, since point B is above the new BP line. This causes monetary expansion, a shift of the LM function to LM , and a move to a final equilibrium at C, with a lower interest rate and exchange rate, and an increased level of national income.
[Figure 3]The Effect of a Decrease in e from 1.0 to 0.85
At the time of writing, the extent to which this demand side analysis will provide an accurate account of the UK's overall adjustment to ERM withdrawal remains to be seen. As before there are some important caveats, arising mainly from economic relationships which are omitted from the model. Firstly, it is uncertain whether UK firms will respond to the improvement in their competitive position by increasing sales of exports (as the IS-LM-BP model assumes) or by taking advantage of the opportunity to raise their prices, in which case the output gains will be much smaller.
Secondly, the prospects for export-led growth for the UK are partly constrained by the buoyancy of the economies of her major trading partners, and with many European economies in recession in the early 1990s, significant doubts surround the UK's growth prospects on this score. In the model, the expansionary effects of the reduction in the exchange rate are substantially less if the 'exogenous' exports parameter x is simultaneously reduced, say from x=300 at t=0 to x=250 from t=1 onwards.
Thirdly and finally, the competitive benefits of devaluation could quickly be dissipated if higher import costs create generalised cost-push inflationary pressures in the domesticeconomy. However, it can be argued that with UK unemployment remaining at extremely high levels, this danger (at least in respect of wage inflation) may be smaller at present than might have been the case on other occasions in the past. With so much spare capacity in the economy, the IS-LM-BP model's key assumptions that output is demand-determined and prices are fixed might not be too far removed from the truth for the UK in the early 1990s. On the other hand, the experience of the last three decades certainly does not inspire excessive optimism in this respect.
The use of spreadsheets can provide a valuable enhancement of students' learning experiences in macroeconomics. The IS-LM-BP model is a particularly appropriate one for student experimentation since it is easy to construct and, in its spreadsheet form, can easily be used to analyse the effects of a wide range of parameter and policy variable changes. From the lecturer's viewpoint, the spreadsheet model is a powerful and flexible pedagogic device, especially for illustrating the dynamic linkages between macroeconomic variables. At one time, the demands placed on students and institutional resources by such an approach might have been prohibitively high. However, computing facilities in higher education have improved considerably, and students on many economics and business-related courses receive some training in the use of spreadsheets (or similar tools). The approach discussed in the preceding sections can be viewed as one way in which the teaching of macroeconomics and spreadsheet skills can simultaneously be improved.
· Other articles on the use of spreadsheets in teaching economics