American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 1 | P a g e THE NEXUS OF GOVERNMENT EXPENDITURE AND FINANCING MECHANISMS: INSIGHTS FROM GHANA Dr. Esi Ama Anaman Department of Banking and Finance, School of Business, University of Education, Winneba, Ghana Abstract: Government expenditure is a pivotal driver of economic growth globally, facilitating infrastructure development and institutional support. However, the financing channels for these expenditures play a crucial role in shaping economic trajectories. This paper delves into the intricate relationship between government expenditure and its financing modes. While conventional wisdom dictates that government spending should primarily rely on direct and indirect taxes, as well as non- tax revenues, many developing nations resort to borrowing to bridge budgetary shortfalls. This paper explores the dynamics of these financing methods and their mutual influences. It sheds light on the challenges faced by governments in balancing their budgets and underscores the imperative of sustainable fiscal policies. Keywords: Government expenditure, Financing channels, Economic growth, Fiscal policies Developing nations Introduction Throughout the world, government expenditure remains an avenue that provides a strong impetus for spurring economic growth and indeed it is the government expenditure outlays in every economy that enable the state to create the necessary infrastructure and the relevant institutional mechanisms to support the multiplicity of economic activities across the spectrum. Whilst it is recognized that the government expenditure is critical in every economy, it must also be noted that such expenditures are usually greatly influenced by the financing channels through which the expenditures are derived. Much as it is true that the trajectory of government expenditure has riposte on the various financing modes, it is also equally an established fact that these financing modes can affect each other .In the literature, there is seem to be a general view that government expenditure must as much as possible be financed from the conventional sources- direct and indirect tax as well as non-tax revenues. However, in the developing world especially, it has become customary to leverage on borrowing modes as a way of meeting the government expenditure levels required in the budget plans as the conventional revenue raising mechanisms always fall way short of the intended targets sufficient enough for the government operations to be pursued seamlessly. Within these contexts, there has emerged a strand of empirical research which seeks to examine the fiscal behaviours of governments and in particular how the availability of the borrowing modes dampens the resolve of the fiscal authorities to be up and doing and maximize revenues. This is well- articulated in the early studies in fiscal behaviours; Griffin (1970), Heller (1975) and Mosley et al.(1987) etc. One important aspect of this discussion centers on the aid effect on the other financing modes and government expenditure itself. According to Osei, Morrissey and Lloyd (2005), studies on the effects of mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 2 | P a g e aid on fiscal behaviour can generally be categorized into those which direct their attention at investigating the effects of aid on the composition of government expenditure and those which in addition to examining the effect of aid on the allocation of government expenditures also assess aid effect on tax effort and government borrowing. In Ghana, just as in a lot of the developing countries, the pressure on successive governments to meet the aspirations of the citizenry has meant that government has to go out of the way to find the needed resources to ensure that programmes and projects are duly executed even against the backdrop of insufficient revenue generated and this situation has persisted for a long time. There are some who believe strongly that this has continued to exist because of the opportunity which is always open for the government to look anywhere to fund its activities even though government could be more prudent in staying reasonably within its revenues limits or aggressively pursuing the much needed tax reforms which could result in enhanced revenue collection. A number of questions thus arise. Does the availability of other government expenditure financing modes encourage government to continue to increase expenditure? Do aid and borrowing dampen tax revenue generation? Do grants and borrowing trigger differential fiscal behaviour by government? Again, how does the availability of the non-tax government expenditure financing modes influence the allocation of government expenditure? Gleaning the literature, it is obvious that contemporary studies in this arena have moved forward the frontiers of knowledge established by the earlier ones, eg Griffin (1970), Heller (1975) and Mosley et al (1987) and Khilji and Zampelli (1994). The most recent study conducted within the Ethiopian context by Mascagni and Timmis (2014) develops a model of fiscal behaviour encompassing tax and non-tax revenues, government expenditure, grants and loans which modifies Osei et al (2005) and Lloyd et al (2009) which include government (capital and recurrent), total tax revenue and domestic borrowing for the former and foreign financing, capital expenditure, recurrent expenditure, tax revenue and domestic borrowing in the case of the latter. In these studies, the researchers did not avert their minds to the fact that the dynamics may not possibly be the same if the tax financing source is disaggregated into direct and indirect tax channels. In other words, in this study apart from categorizing aid as grants and loans, we also include direct and indirect tax financing as separate variables. This is because we believe that aid and borrowing may not necessarily have the same effects on direct and indirect taxes. Thus the main difference between the present study on one hand and that of Osei et al (2003) and other previous but related studies on the other hand is that it we introduce the hypothesis that the responses of direct and indirect taxes respectively to borrowing-both external and domestic are different and also have the benefit of current data for the analysis to determine whether prevailing circumstances deviates from Osei et al(2003). The rest of the paper would be arranged in the following manner; Section II is devoted to examining the fiscal policy environment, trends in fiscal management and borrowing by the government of Ghana over the years. In section III, we proceed to discuss the theoretical and empirical issues relating to fiscal behaviour by government especially focusing on aid and borrowing and their effects on government fiscal management. Section IV sets out the econometric approach and a brief description of the data set for the empirical analysis whilst Section V reports the results of the data analysis and proceeds to discuss them. Finally, section VI covers the synopsis and conclusions from the study. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 3 | P a g e Section IIː Trends in Fiscal management in Ghana When one does a careful study of fiscal policy in Ghana, one can identify clear, distinct periods of unique fiscal behaviours? In the main, the periods the early 1960s to the late 1960s, 1969-1972, 1972-1983, 1983-1991 and from 1992 to the present can be associated with peculiar fiscal behaviours though in some of the periods the fiscal management approaches appear similar. In the sixties, with the emergence of the country from colonial rule there was an urgent need for the government to put in place structures of state and build critical infrastructure like educational and health institutions while also embarking on rapid industrialization and modernization and as such, government committed massive public expenditures into achieving these objectives. During this period, a good chunk of the expenditures were financed from domestic sources with very little coming by way of aid inflows. The succeeding period however saw a modification of fiscal behaviour as government substantially disengaged from the previously pervasive role of the government in the economy, in line with the philosophy of the people in authority at the time and by virtue of the programme that they entered into with the Bretton Woods institutions, government at the time embarked on privatization of a good number of the state enterprises. Osei et al(2003) submit that from the 1960s to early 1970s , aid inflow was relatively insignificant and constituted about 2% of GDP and roughly around 12% of all revenues available to government. In the middle to the late 1970s, there was a shift in the behaviour of the government as government activities were driven essentially by monetary expansion through borrowing from the Bank of Ghana as domestic revenues sharply reduced on account of the decline in the real side economic activities precipitated by inappropriate policies introduced by the then military rulers coupled with adverse economic and external trade climate. The situation was compounded by the repudiation of loans which had been contracted by previous governments leading to the virtual drying up of the foreign aid inflows. In early 1980s, even though the country had returned to constitutionalism, the country continued to suffer from the decline in economic activities as result of the deterioration of the macroeconomic environment. According Durdonoo’s (2000) calculations, taxes on income and property fell from 2.8% of GDP to a mere 0.98% of GDP in 1983 whilst tax revenue from domestic activities was down to sub one percent in 1983 from approximately 5% of GDP. Proceeds from international transactions also dropped from 12% in 1970 to 2.7% in 1983.The precarious revenue situation in the country is illustrated by Osei et al(2003) when they intimate that overall the tax levels took a nosedive between 1970 and 1983,plummeting from a high level of about 700 million USD to 160million USD. The fiscal situation in the country however improved dramatically after the Economic Recovery Programme (ERP) was launched. Indeed it is estimated that between 1983 and 1998, tax revenue collections shot up in dollar terms to 1.3 billion USD representing a more than six fold increase of the 1983 level. Generally speaking, total government revenue is measured in some calculations to have increased twenty-six times between 1983 and 1990. Osei et al (2003), suggest that since this period was largely marked by good amount of aid inflows, it appears that the aid flows did not undermine government's tax revenue mobilization. During the period whilst tax revenue and aid inflows were increasing, government expenditure also continued to increase though at a slower pace, on account of the some of the ERP measures which had been introduced to stem rapidly increasing government expenditure experienced in the period before ERP. The post 1992 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 4 | P a g e period has generally been characterized by rapidly expanding government expenditure largely fueled by the people across the country making demands on politicians and indeed some actually using the provision of certain projects and infrastructure as a tool for cajoling and blackmailing functionaries of government. However, revenue mobilization during the period has not kept pace with government expenditures and the peoples aspirations for that matter. Indeed for a long time now total government revenue is consumed largely by payment of emoluments, statutory obligations and then interests payments, which itself is a product of the increasing imperative for borrowing by government. The argument that the political structure in Ghana has tended to reinforce fiscal behaviours by successive governments in the fourth republic is amplified when one considers Ghana’s fiscal position in election years. Election cycles have generally exacerbated the problem and this is evidenced by the fiscal deficits which were recorded in the years 2008, 2012 and 2016 respectively. One major development which has also to a great extent influenced fiscal behaviours especially post 2012 has been the reclassification of Ghana as a middle income country .This has restricted the country's access to concessionary loans and grants and compelled governments to syndicate relatively expensive loans from the international commercial markets on account of the fact that Ghana's tax to GDP ratio is woefully below the average middle income levels. Indeed it is very instructive to note that in the West Africa sub region, Ghana's tax collections as a percentage of GDP is the lowest. Against this background, the issuances of Euro bonds have become an important feature of government strategy for financing projects and programmes of government as missing revenue targets have become a constant feature of fiscal management in Ghana. Section III ːTheoretical, conceptual issues and empirical underpinnings Fiscal policy formulation is one of the basic functions of every government in the sense that it primarily involves the strategies that governments use to raise income to be able finance government's activities. In the main, most governments rely on revenues generated from taxation as the most reliable source of income. However, in most parts of the world particularly the developing world because of the demands on government to ensure rapid development and the exigencies of the time, they are unable to stick to the incomes available to them through taxation and therefore have to resort to other means of financing their programmes and projects. These come in the form of foreign aid- loans and grants and domestic borrowing. According to Njeru (2004), one of the most critical issues which has been a subject of debate by economists in this area of research is whether or not the aid process is undermined by the ability of the aid receiving country to alter the their spending patterns to subvert the sectoral distribution of expenditure for designated projects. The general contention is that the ability of the recipient country to reallocate the aid can usually affect the intended economic performance envisaged under the aid structure. This is particularly the case when aid earmarked for developing critical infrastructure in a given economy is diverted into financing government consumption like catering for emoluments of workers and buying goods and services for government machinery rather than creating the required infrastructural overheads which then provide the necessary platform for increasing the level of economic activities. This is what economists usually refer to as aid fungibility. This is reinforced by Bwire et al (2017) who contend that fungibility arises when aid recipients do not use the aid for purposes for which they were given by the donors. Thus in many respects, a lot of the developing countries employ mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 5 | P a g e resources from aid to able to be able to deal with the deficits usually associated with their budgetary processes (Devarajan et al, 1998, Ali et al , 1999). These views are very replete in the fiscal response studies. For example Matins (2007) asserts that one of the most fundamental issues which relate to the effectiveness of aid is how aid influences the government fiscal accounts. In particular, Martins (2007) stresses that one critical pillar of the fiscal response studies is assessing how the aid itself is allocated between the various expenditure channels, the way it affects tax effort and then its effect on fiscal balance and debt sustainability. This is view is reinforced by Mavrotas (2002) who stresses that since aid is given to a government, its impact on the overall economy is contingent on fiscal behaviour of the government. From the perspective of Mascagni and Timmis (2014), aid is usually a more politically expedient and convenient source of revenue and therefore has the tendency to discourage tax effort which in the literature is characterized as tax displacement. They however stress that this argument is stronger in respect of grants than loans because of the obvious fact that loans require future payments whereas grants do not. Mascagni and Timmis (2014) put forward another dimension of the aid–revenue debate which is that rather undermining the revenue efforts, aid may actually help strengthen tax administration and improve tax policies. Again, it is argued that if aid is utilized properly and effectively it may promote economic activities, expand the economy and by that increase tax yields from the economy. In the view of Njeru (2004), aid inflows into the developing countries has tended to create an ominous dependency mentality which seem to affect their economic performances and the absence of such funds greatly affect their budgets, usually coming with their attendant consequences. This is echoed by Feyzioglu et al (1999) who posit that aid dependence is something which has widespread ramifications for countries. There is an also another dimension of the aid debate which is canvassed by Martins(2007) .In his estimation apart from the fact that aid is sometimes used to offset domestic debts, it can trigger off extra government expenditures especially in aid funded projects which require some maintenance and recurrent expenditure. Again aid programmes and projects which require counterpart funding may in reality also further put pressure on government's already overstretched finances and thus lead to mounting deficits. This scenario is what McGillivray and Morrissey (2000) describe as aid illusion. Having regard to the fact that foreign aid may be associated with some challenges; the other viable alternative is borrowing from domestic sources to be able to undertake the necessary government activities. However, this avenue also comes with its own problems. One of the challenges that this poses is that it leads to a situation in which government enters the credit markets to compete with private entities for the available funds, a situation which generally inhibits the growth of privately engineered economic growth in an economy. This can in many respects also affect tax mobilization. Aside of these issues ,it is often argued that in a lot of the developing countries, excessive reliance on borrowing modes to enable governments meet its commitments in terms of delivering the required services has invariably led to compounding debt servicing obligations and thereby constricting fiscal space as the piling of debts both internally and externally have tended to increasingly impose severe servicing and payments obligations on the government thereby limiting what the government can achieve within its resource envelop. Studies in fiscal response has its origins in the 1970s starting with Heller(1975) who used a utility based government fiscal behaviour function to show that the aid process has effects on how governments manage their fiscal operations . Despite the vast array of mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 6 | P a g e research in this area, results from these studies have largely been inconclusive. According to Mascagni and Timmis (2014), this situation may be due to the fact that various studies adopted different methodologies and contexts. In his study, Njeru (2004) assessed the impact of foreign aid on public expenditure in Kenya and based on government welfare utility function specified government expenditure related to aggregated government revenues from tax and domestic borrowing sources, programme aid and project aid. The dynamic analysis indicated that aid does not affect government expenditure whilst it is also established that government is able to divert aid funds into government consumption expenditure. A similar study by Osei et al (2003) modelled the fiscal effects of aid in Ghana by particularly employing a dynamic impulse response functions. Using the government utility maximization approach, two variants of the empirical model were specified; these are aggregate government expenditure ,domestic borrowing ,total government tax revenues and aid finance on hand and government capital expenditure, government consumption expenditure, domestic borrowing, total tax revenue and foreign aid. In the analysis, it is established that there are co-integrating relationships in both models. Results also showed that in both models, aid finance and domestic revenues are in long run negatively related to domestic borrowing whilst government expenditure whether aggregated or disaggregated positively influences domestic borrowing .Another important finding that issues from Osei et al (2003) is that aid in Ghana over the study period has generally been used to replace domestic borrowing as a method of financing government projects and programmes. The work of Martins (2007) also explores further the aid-fiscal behaviour nexus within the context of the Ethiopian economy and actually separates aid into two components-loans and grants based on the premise that fiscal response by government to them may be different. The conclusions from the estimations are that whilst aid finance positively affects total government expenditure, its effects on government consumption expenditure is less pronounced and that external borrowing has a bigger impact on public investment than grants. Another important finding from this study is that aid finance undermines domestic revenue mobilization. In his contribution in the fiscal response and effectiveness of aid studies, Mavrotas (2002) introduced a categorization of foreign aid into project aid, programme aid, technical assistance and food aid and based on the popular utility maximization approach obtained results which affirm that aid may be fungible. The study by Mascagni and Timmis (2014) also dealt with the fiscal effects of aid in Ethiopia employing the co integrated vector autoregressive model based on the conventional Heller utility maximization function. Their model encompassed total government expenditure, tax and non-tax revenues, grants and loans .In the long run , government expenditure was established to be related to domestic revenue and foreign aid; there is a positive relationship between tax revenue ,grants and loans. In the short run too, government expenditure is established to be influenced positively by both grants and loans whilst the equation for tax shows that non-tax revenue, grants and loans are all positive determinants. The, loans variable is also impacted positively by non-tax revenue but negatively by tax revenues. The most recent study in this area, authored by Bwire et al (2017) also sought to examine fiscal reforms and the effects of aid in Uganda employing a dynamic analysis and to test whether aid flows lead to a full or less than a full change in government expenditure, determine if aid displaces tax effort as well as ascertain mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 7 | P a g e whether aid and domestic borrowing are substitutes in fiscal management in Uganda. Their analysis uncovered three co integrating equations for government expenditure, revenue and aid and found that in the long run aid leads to increased tax effort and public spending but a reduced domestic borrowing. Section IVː Empirical model According to Osei et al (2003), there are two broad approaches adopted in the literature to examine the fiscal effects of aid. The first approach is the fungibility studies which attempt to assess aid effects on the spending patterns of the government and the other method which seeks to integrate revenue variables into a government utility function to determine the overall impact of the aid process on the fiscal behaviour of the government which they call the fiscal response models (FRMs). Since the latter is more comprehensive in its outlook, it is more popular in the literature and has been adopted in most of the recent studies. This approach is based on the seminal work of Heller (1975) which posits government allocating revenue among the different expenditure streams but subject to some budget constraints. In the model, government expenditure is usually categorized into government consumption and capital expenditure whilst government derives its income endogenously from conventional taxation sources and domestic borrowing. However, in these models, foreign aid is defined as an exogenous source of revenue which modifies the government budget constraints; even though it is assumed not to be relevant in the utility function of the government since it is not defined as one of the variables for which targets are set. Against this background, Osei et al (2003) set the maximum unconstrained value of the utility function represented by α0 as a quadratic expression defining a loss in the form below; U=α0–α1/2(GK-GK*)2-α2/2(GC-GC*)2-α3/2(R-R*)2-α4/2(D-D*)2 (1), where GK*, GC*,R* and D* are exogenous target values of government capital expenditure ,government consumption expenditure ,total government revenue and government borrowing from domestic sources. The above equation is thus maximized subject to the following budget constraints, GK = (1-ρ1)R + (1-ρ2)F +D (2) and GC=ρ1R+ ρ2F (3) Where equations (2) and (3) are disaggregated equations derived from the total government expenditure constraints, of the form, GK +GC = R+F+D ( 4) From the above equations, it is taken that ρ2 represents the fraction of aid which is diverted into financing government consumption ;in other words the extent of the fungibility of aid .The implicit argument underlining this formulation is that when foreign aid is received , it is meant for capital investment .However , as the aid comes into the economy, a part of it is channeled into financing recurrent expenditure which means that mathematically, ρ2=0 ex ante but this according to Osei et al (2003) is not in the real world realistic because aside of directing resources into investments in the economy, foreign aid sometimes finances certain components of government consumption ,particularly in the social sectors especially education and health, hence ρ2≠ 0 is an unrealistic assumption but ρ2> 0 at most times is the most realistic assumption to make. This situation occurs especially when aid comes in in the form of budgetary support or even strictly as aid funded project in an economy. With the inherent limitations of this approach, Franco-Rodriguez et al (1998) modified mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 8 | P a g e the approach by defining a utility function such that foreign aid is interacted and integrated directly into the function thereby making aid endogenous. The argument put forward is that generally governments define targets for aid flows and this tends to influence their fiscal behaviour. As a result, the quadratic utility loss function expressed in (1) becomes U =α0– α1/2(GK-GK*)2- α2/2(GC- GC*)2- α3/2(R-R*)2 –α4/2(F-F*)2 -α5/2(D-D*)2 (5) whilst the constraining function becomes GC ρ1R + ρ2F + ρ3D (6) since external flow of funds tend to influence how resources are allocated among competing needs. In this current paper, we further redefine (5) as U =α0– α1/2(GK-GK*)2- α2/2(GC- GC*)2- α3/2(DT-DT*)2 – α4/2(IT-IT*)2 -α5/2(F-F*)2-α6/2(D-D*)2– α7/2(Gr- Gr*)2 (7) Subject to GC ρ1DT + ρ2IT + ρ3F+ρ4D +ρ5Gr (8) The implicit meaning of the above is that both direct and indirect sources of revenue are endogenously determined in addition to the other sources of revenue. In this formulation, we separate F, external borrowing from Gr, grants because in the literature, it is argued that most governments treat loans differently from grants which are not to be paid back. Though Franco-Rodriguez et al (1998) provided an improvement of the earlier fiscal response models (FRMs), they did not address the methodological challenges that most of the earlier studies were fraught with. Osei et al (2003) therefore in their study changed direction to the new vector autoregressive (VAR) approach which in their view provided the means to go round the existing problematic methodological frameworks whilst making it easier to define the dynamic linkages between the various components of the budget. Building upon Osei et al(2003), M'Amanja et al (2005),Martins (2010),Bwire et al(2017) and Mascagni and Timmis(2014), we specify two variants of the VAR model involving aggregate and disaggregated government expenditure models belowː (GE, DT, IT, Db, Fb, Gr) and (GK, GC, DT,IT, Db, Fb, Gr) respectively where GK is government capital expenditure, GC is the government consumption expenditure, IT is indirect tax revenue ,DT defines direct tax revenue, Db represents domestic borrowing ,Fb is used for external borrowing whilst Gr is grants obtained from various external sources and finally GE defines aggregate government expenditure. In these models above, the application of the VAR allows us to determine whether the variables are in the long run are dynamically related whilst at the same time providing useful information about the short run properties of the models. Generally an orthodox VAR model is defined as a dynamic system in which all the variables are endogenously determined and each of them is represented as a function of its own lags and the lags of the other endogenous variables. The advantage from this, according Blanchard and Peroti (1999) is that the system assumes a priori there is no direction of causation among the variables of interest. Mathematically, we define our VAR (k) as Xt=φ1Xt-1 + φ2Xt-2 +φ3Xt-3 +…..+φkXt-k+ πRt + t, t=1,2,----,n. (9) mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 9 | P a g e In the above, Xt is defined as a (m × 1) vector of non-stationary variables whose order of integration is one and which are jointly determined, whilst the Rt is also a vector of deterministic variables of dimension, (p× 1 ). The coefficients π and φ which are to be estimated are matrices of the dimension (m p) and (m m) respectively whereas the disturbance term t is a vector of dimension (m 1) and k is the lag length of the system. Using the Johansson (1991) approach, the general VAR can be transformed into an error correction model, usually referred as a restricted VAR of the form ∆Xt=α+ψXt+τ∆Xt-1+……..+τk-1∆Xt-k+1+ ∈t,t=1,2,-----,n. (10) In this expression, we use the τs to define the short run characteristics of the variables. Specifically the coefficients of the lagged dependent variable represent the feedback in the system whilst the coefficients of the other endogenous variables in the system define the pass through effects of these variables on the dependent variable. The matrix of coefficients ψ represents the long run equilibrium relationships among the variables of interest in the system. We start the analysis by examining the stationarity properties of the variables. This is important because in empirical analysis, most macroeconomic variables have been found to be non-stationary as result of their time dimensions and as a result prejudice and distort estimations. This thus makes it imperative for the non-stationary properties to be dealt with. In the words of Thomas (1993),if a variable is stationary, it means that the time path traced by the variable is stable. In other words, a series is said to be stationary when it has a spectrum which is finite but non-zero at all frequencies. Mathematically determining the stationarity of a series Yt involves finding whether the equation Yt = α0+α1t +α + ut (11) follows an AR process. Typically, assessing the stationarity properties of variables involves testing the following hypotheses; H0ː The series has unit roots ,H1 ːThe series has no unit roots. In the conventional VAR system, the order of integration of the variables allowed is one meaning that the each variable in the system must attain stationarity after first differencing. Beyond examining the stationarity status of the variables, we employ the Johansson approach to test for co integration; that is to ascertain whether there exists a linear combination of the variables which is also stationary. According to Anaman et al (2017),co integration is the statistical implication of the existence of a long run equilibrium relationship between economic variables. Soli et al (2008) also characterize co integration as representing the tendency of variables to drift together over time. The obvious advantage in the Johansson approach over the other methods of determining long run equilibrium relationships is that it makes it possible to uncover more one co integrating vector at a time. To proceed with this, we test, H0ːthere is no long co integrating vector in the system, as opposed to H1ː At least one co integrating vector exists in the system. In a VAR system, a major requirement is that for variables to be co integrated they must have the same order of integration. This is underscored by Enders (1995) who emphasizes that for variables to be co integrated they must be integrated of the same order and have a linear combination of residual sequence which is stationary . mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 10 | P a g e In order to extract more information from the system and sufficiently understand the dynamic relationships among the variables in our system, we undertake impulse response analysis. The importance of the impulse response analysis is buttressed by Osei et al (2003).According to them, when the interrelationships that characterize economic systems are considered, it is always more informative to undertake an impulse response analysis especially when the analysis involves uncovering short and long run relationships within a given system. Osei et al (2003) assert that the advantage that the impulse response analysis has is that it captures the net effect of both the direct and indirect impact of a shock, not only in the long run but also at all periods after the shock has been transmitted. Johnston and DiNardo (1997) underline the relevance of the impulse response function by intimating that it traces the chain reaction or the knock-on effects arising from one standard deviation perturbation in one innovation in the system over time on the other variables in the system granted that no other shock affects the system afterwards. Impulse response functions can thus measure both the current and future values of the given endogenous variable to one standard deviation shock in one of the innovations. Lutkepohl and Rimmers (1992) also reinforce the importance and suitability of the impulse response in a dynamic analysis. Generally, the impulse response function can be defined as the moving average representation of our equation (9) ,expressed as Xt πRt-1 (12), where, the As are of dimension (m m) Apart from the impulse response analysis, we employ the forecast error variance decomposition from our VAR model to ascertain and predict the most important innovation for each endogenous variable along the entire time horizon. This will enable us to identify which variable is most relevant in achieving a given objective. According to Bhasin (2004), in a VAR model, variance decomposition is usually employed to isolate the innovations of the endogenous variables into the portions which can be attributed to own innovations and that which are due to innovations of other variables in the system and in doing so we recourse to the Cholesky method based on Sim's recursive approach. Data Set For the purposes of this study, we employ annual series for all the variables from 1978 to 2017 .The variables were largely extracted from the World Bank Databases and supported with data from Ghana Statistical Service(GSS) and the Bank of Ghana. Results of Data Analysis Test for stationary (Unit roots tests) In the tables below, we report the results of the stationary tests. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 11 | P a g e Table1ːUnit root tests of log levels of variables Variable (log levels) ADF test Statistic Prob level Phillip Perron test statistic Prob level ldb -2.298031 0.1778 -1.917340 0.3212 ldt -1.682815 0.4318 -1.597301 0.4744 lfb -0.413932 0.8968 -0.404310 0.8985 lgc -0.141300 0.9375 -0.072070 0.9455 lge -1.130286 0.6941 -1.120580 0.6980 lgk -0.724807 0.8286 -0.710912 0.8322 lgr -1.946252 0.3085 -1.946252 0.3085 lit -2.194319 0.2115 -2.057338 0.2623 Sourceː Author’s calculations using E Views Table 2 ː Unit root tests of first differences of variables Variable (first differences) ADF statistic Prob. level Phillips Perron Statistic Prob. level dldb -4.411600 0.0012 -4.425672 0.0011 dldt -5.736767 0.0000 -9.443833 0.0000 dlfb -6.196007 0.0000 -6.195848 0.0000 dlgc -5.475900 0.0001 -5.466520 0.0001 dlge -5.076895 0.0002 -4.944660 0.0003 dlgk -5.107552 0.0002 -5.032008 0.0002 dlgr -6.775472 0.0000 -6.799061 0.0000 dlit -7.863574 0.0000 -9.818331 0.0000 Sourceː Generated from E Views estimations From tables 1 and 2, we infer that all variables are non-stationary at log levels but are stationary at first differences .This means that the order of integration of all variables is one. We proceed to determine the optimal lag for the disaggregated and the aggregated models respectively. For the disaggregated model, we determine whether or not there is first or higher order serial correlation in the initial model by performing the autocorrelation LM test. The test results are presented below Table 3 Included observations: 37 Lags LM-Stat Prob 1 39.13263 0.8424 2 48.47734 0.4942 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 12 | P a g e From the results generated , we fail to reject the null hypothesis that there is no serial correlation in the model meaning that the model is can be correctly specified by using the first lags of all variables. We proceed to corroborate the above finding by determining the optimal lag structure of the model using various criteria. The table below shows the selected optimal lag structure using various criteria for the disaggregated model. Table 4 Included observations: 36 Lag LogL LR FPE AIC SC HQ 0 39.11941 NA 3.96e- 10 - 1.784412* - 1.476505 -1.676944 1 70.04318 48.10365* 1.14e- 09* - 0.780177 1.683075* 0.079564* 2 114.0449 51.33531 2.08e- 09 - 0.502493 4.116104 1.109521 3 170.9768 44.28037 3.70e- 09 - 0.943154 5.830788 1.421133 * indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) The results confirm that appropriate lag to be used in the analysis is one considering that four out of the five criteria settle on lag one. In the case of the aggregated model, the test for 1st and 2nd order serial correlation is presented in the table below. Table 5 Null Hypothesis: no serial correlation at lag order h Sample: 1978 2017 Included observations: 38 Lags LM-Stat Prob mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 13 | P a g e 1 47.48022 0.0955 2 50.66403 0.0533 From these statistics, we conclude therefore that in the aggregated model, there is evidence of serial correlation in the residuals at lags one and two so we proceed to determine the optimal structure for the model. The following table shows the selection analysis. Table 6 Included observations: 36 Lag LogL LR FPE AIC SC HQ 0 26.83531 NA 1.27e-08 -1.157517 - 0.893598* - 1.065402* 1 59.96044 53.36827 1.53e-08 - 0.997802 0.849636 -0.352997 2 101.9825 53.69481 1.30e-08 -1.332359 2.098598 -0.134863 3 156.1070 51.11764* 7.91e- 09* - 2.339279* 2.675197 - 0.589093 * indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) From the results shown in the table above, we firmly conclude that the optimal lag for the aggregated model is three based on the different criteria. Having completed the tests for stationarity and optimal lag structures for the two models we then enter the log levels of the variables in the two models into the Johansson test for co integration, the results of which are presented in the tables below. Table 7 Johanssen test for co integration for the disaggregated model Series: lgc lgk ldt lit lfb ldb lgr Unrestricted Cointegration Rank Test (Trace) Hypothesized Trace 0.05 No. of CE(s) Eigenvalue Statistic Critical Value Prob.** None * 0.716509 146.2219 134.6780 0.0087 At most 1 0.623665 98.31997 103.8473 0.1098 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 14 | P a g e At most 2 0.424277 61.18352 76.97277 0.4275 At most 3 0.318226 40.20262 54.07904 0.4603 At most 4 0.256078 25.64646 35.19275 0.3620 At most 5 0.187746 14.40532 20.26184 0.2625 At most 6 0.157301 6.503534 9.164546 0.1553 Trace test indicates 1 cointegratingeqn(s) at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level **MacKinnon-Haug-Michelis (1999) p-values Sourceː Generated from E-Views. From the results presented above, we reject the hypothesis that there is no co integration in our series in favour of the alternative hypothesis that there is one co integrating equation in our model. Using the un-normalized coefficients, we derive the long run equation for government consumption expenditure below by normalizing on government consumption expenditure. Table 8 Long run equation for government consumption expenditure LGC LGK LDT LIT LFB LDB LGR C 1.000000 0.265024 0.945707 4.311506 - 2.071946 0.967689 -1.537559 38.95029 (0.32764) (0.58921) (0.72563) (0.50301) (0.43209) (0.30392) (6.55926) Sourceː Generated from E-Views. From the long run equation, we observe that government capital expenditure, direct taxes, indirect taxes as well as domestic borrowing have negative effect on government consumption expenditure with about 27%,95%, 431%and 97% impacts respectively on government consumption expenditure with a 100% increase in each of the variables. However, in the long run , borrowing from abroad and grants are financing sources which have positive impact on government consumption expenditure. Specifically, a 100% increase in external borrowing in the long run triggers a little over 207% increase in government consumption expenditure whilst a 100% increase in grants also leads to a 154% upswing in government consumption expenditure. In the table below, we present the results of the tests for co integration in the aggregated model. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 15 | P a g e Table 9: Test for Co-Integration in the Aggregated Model Series: LGE LDT LIT LFB LDB LGR Hypothesized Trace 0.05 No. of CE(s) Eigenvalue Statistic Critical Value Prob.** None * 0.642654 127.6208 103.8473 0.0006 At most 1 * 0.606512 88.51696 76.97277 0.0050 At most 2 0.493073 53.07422 54.07904 0.0613 At most 3 0.275454 27.25744 35.19275 0.2761 At most 4 0.199710 15.01346 20.26184 0.2256 At most 5 0.158281 6.547753 9.164546 0.1525 Trace test indicates 2 cointegrating eqn(s) at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level Sourceː Output generated by author from E-Views using data In the table above, we test the null hypothesis that there is no co integrating relationship in our series against the alternative hypothesis that there is at least one co integrating relationship. From the table above we fail to accept the hypothesis that there is at most one co integrating relationship but fail to reject the null hypothesis that there are most two co integrating vectors in our model. This thus means that in our series, we can uncover two co integrating relationships. In the table below, we present the first co integrating equation from our model. Table10. Long run function for aggregated government expenditure LGE LDT LIT LFB LDB LGR C 1.000000 7.033729 18.66572 -8.250564 -0.279982 -6.581200 259.1778 (2.36778) (3.53366) (1.53188) (1.71301) (1.37989) (30.0200) Sourceː Generated by author using E-Views estimation. From the results, we determine that in the long run, direct and indirect taxes negatively impact on government expenditure whereas external borrowing, domestic borrowing and grants exert a positive effect on government expenditure. The estimated negative long run impacts of direct and indirect taxes on government expenditure are respectively 7.03 and 18.67 units as each of these increases by a unit. On the other hand, a unit increase of each of external borrowing, domestic borrowing and grants leads to about 8.3, 0.28 and 6.58 units increase in government expenditure. We derive the second co integrating equation from the un–normalized co integrating coefficients by normalizing on external borrowing .We thus derive the long run equation for external borrowing in the form below; Table 11.Long run equation for external borrowing LGE LDT LIT LFB LDB LGR C -0.346960 0.850657 0.668850 1.000000 -1.277610 0.027437 - 13.132143 Sourceː Output generated by author based on E-Views estimations. From the table, we define the long run equilibrium relationship between external borrowing and the endogenous variables. In this relationship, we observe that government expenditure and domestic borrowing exert positive effects on external borrowing meaning that in the long run an increase in both mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 16 | P a g e government expenditure and domestic borrowing lead to an increased external borrowing. Generally from the estimation results, a 100% increase in government expenditure leads to about 35% increase in external borrowing whilst a 100% increase in domestic borrowing calls forth a whopping 128% increase in external borrowing. On the other hand, direct taxes, indirect taxes as well as grants expectedly all impact negatively on external borrowing. More specifically, in the long run a 100% increase in each of direct taxes, indirect taxes and grants precipitates about 85%,67% and 3% decline in external borrowing. Using tables 8,10 and 11 , we derive the error correction terms ect1,ect2 and ect3 respectively which are entered into the short run models to determine the short run effects of each of the endogenous variables on the other endogenous variables in tour system. Short Run Relationships Proceeding with our analysis, we estimate the short run/error correction models for the government consumption expenditure, aggregated government expenditure and external borrowing (the estimates are provided in the appendices). In these models, we observe the signs of the error correction terms are all negative meaning that the behaviours of the short run equations are in line with the theory that once these are co integrated then there is a tendency for each of them to be moved towards the desired equilibrium position.; that is each system is eventually drawn towards the equilibrium time path when there is a deviation from their expected long run position. Of the three models, the equation for aggregated government expenditure is estimated to have the fastest return to its equilibrium time path after a deviation with a speed of adjustment of about 81% per period. This followed by the equation for external borrowing with a speed of adjustment of about 43% per period when it deviates from the equilibrium .The government consumption equation however has about 20% of its deviation from the long run corrected in each period. In the general government consumption equation, our short run estimates show that the government consumption expenditure is significantly impacted by a feedback, growth in government capital expenditure, direct taxes, domestic borrowing and grants. Their contemporaneous effects are estimated at -0.701589, 0.564860, 0.3202703,0.223057 and 0.203551 respectively. This shows that previous period government consumption expenditure tends to have a negative impact on current government spending on consumption. It is also noticed from the estimation that the previous government capital expenditure has a positive effect on current government consumption expenditure. This result contrasts with Osei et al (2003). In actual terms, from the results, a 1 unit increase in the previous period government consumption expenditure triggers about 0.56 unit increase in current government consumption expenditure. Lastly a previous increase in grants precipitates an increase in current government consumption expenditure with a 100% previous increase in grants leading to a 20% increase in the current values of general government consumption expenditure. In the aggregated government expenditure function, just as is witnessed in the consumption expenditure equation registers a negative feedback with a magnitude of 0.134650 per unit increase in government expenditure. The only difference is that in the case of the aggregated government expenditure function, the feedback comes from the third period. Also, growth in direct taxes and indirect taxes respectively exert positive and negative effects on government expenditure with contemporaneous impacts of about 1.15 and 0.40 when there is a unit increase in each of them. Thus the dynamic effects of domestic revenue from these results appear mixed, and therefore do not fall wholly in tandem with the finding of Njeru (2004). mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 17 | P a g e In relation to external borrowing, the short run behaviour is explained by growth in aggregated government expenditure, direct taxes, indirect taxes as well as grants. In the estimates, it is seen that a 100% increase in government expenditure in the first period expectedly leads to about 35% increase in current levels of external borrowing. From the short-run equation, it is also realized that in the third period, increased direct taxes leads to a decline in external borrowing. From the results, a 100% increase in direct taxes tends to lead to about 92% decline in external borrowing which falls in line with expectation that increased domestic revenue mobilization leads to reduction in dependence on external sources of financing projects and programmes. The indirect tax variable, in the period also elicits a negative response from external borrowing. The measured effect, significant in the first period is even bigger in magnitude than that of direct taxes. In real terms, a 100% growth in indirect taxes precipitates over 138% decline in external borrowing. The effect of growth in grants on external borrowing is felt in two periods-the first and third periods and in both periods their impacts are positive. In the first period, a 100% growth in grants tends to increase external borrowing by about 105% whereas in the third period, the effect is smaller at 0.31 unit’s growth in external borrowing with respect to a unit increase in grants. One major position which is dominant in the literature that we wanted to verify was whether or not the availability of other sources of financing government activities dampens tax effort. In the disaggregated government expenditure model, we are unable to substantiate the hypothesis that external financing tends to stunt domestic mobilization of revenue. Our regression results indicate that the impact of external borrowing and grants are positive and negative respectively. Thus for 100% increase in external borrowing, we experience about 33% increase in direct taxes but the same amount of increase in grants precipitates a 4% decline in direct taxes. For the aggregated government expenditure, the story is similar that is positive and negative in respect of external borrowing and grants respectively. The impacts of external borrowing and grants on indirect taxes are mostly insignificant except in the aggregated expenditure model in which growth in external borrowing triggers a decline in indirect tax yield. These findings are partly consistent with Mascagni and Timmis (2014) who discovered positive but significant impacts of grants and loans on the tax revenue variable.In the aggregated model, growth in external borrowing rather than leading to a decline in growth in direct tax mobilization actually triggers an increase. This finding coincides with Osei et al (2003). From the estimated equation, a 100% growth in external borrowing in the second period precipitates about 42% growth in direct taxes in the current period. However its estimated effect on indirect taxes is negative .The estimates indicate that a 100% increase leads to about 24% decline in indirect tax. The effect of domestic borrowing variable on the revenue variables- direct and indirect tax is very interesting. In the aggregated government expenditure models, we notice a negative impact of domestic borrowing on both direct and indirect tax variables. However whilst its effect is negative and significant with respect to growth in direct taxes, the measured impact is not significant in the case of indirect tax. From the estimated restricted VAR, a 100% growth in domestic borrowing elicits about 42% decline in direct taxes. On the other hand, for the disaggregated government models, the effect of domestic borrowing on both direct and indirect taxes is in line with the results for the aggregated model , meaning that an increased growth in domestic borrowing also impacts negatively on both direct and indirect taxes. Finally the effects of grants on the revenue channels-both direct and indirect taxes are estimated to be negative .Whilst its effect on direct taxes are significant that on indirect tax is insignificant. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 18 | P a g e Results of Forecast Variance Decomposition In line with conventional dynamic analysis, we proceed to do a variance decomposition of the residuals of the variables and the results can be gleaned from the appendices of the paper. In dynamic analysis, variance decomposition is particularly very relevant in determining how much of the variation in a given variable can be traced to own innovations and innovations from other variables. The decompositions are performed on the basis of the aggregated and disaggregated government expenditure and in consonance with the Cholesky approach which ensures that the decomposition is carried out maintaining the ordering of the variables just as pertains in the co integration test as well as the error correction estimations. In the aggregated government model, we determine the most important innovations for attaining a particular objective for the various variables the aggregate government expenditure, direct taxes, indirect taxes, external borrowing, domestic borrowing and grants. From the results generated, it is clear that in respect of government expenditure, from the second period, growth in grants assumes a very important position in accounting for over 50% of the behaviour of the government expenditure variable. In the long run, it accounts for close to 70% of the movements of the government expenditure variable. For direct taxes, own innovations are largely responsible for its variations in the short to the long term accounting for over 85% of its movements. The next most variables are indirect taxes and growth in government expenditure which between them from the short to long term explain more than 30% of the movements in the direct tax variable. From the variance decomposition of the indirect tax variable, its movements in the short term are dominated by own innovations and that from direct taxes. However in the medium to the long term the most important variable that influences movements in indirect tax is growth in grants. In respect of growth in external borrowing under the aggregated model, in the short to the medium term, own innovations are largely responsible for its behaviour though in the long run, growth in grants assumes the most dominant position accounting for just over 39% of variations in external borrowing. For domestic borrowing, in the short to the medium term, its variations are explained mainly by own innovations and that from external borrowing accounting for over 90% to about 30% between them. In the long term, however, growth in grants becomes the most dominant as it caters for over 54% of variations in the domestic borrowing variable. For grants, its own innovations are most dominant in explaining its movements from about 51% in the first period to over 60% in the tenth period. In the short to the medium term, however, the growth in external borrowing is second most important innovation which affects movements in grants. We now consider the forecast error decomposition in the disaggregated government expenditure model. From the derived results, we observe that from the short to the long term, the important variable that explains the behaviour of government consumption expenditure is own innovations which constitutes 100% to 60% of its movements from the short to the long term. It is followed in terms of significance by the innovations due to government capital expenditure. In respect of government capital expenditure, the movements are mostly explained by own innovations and that emanating from government consumption expenditure. The movements in direct taxes are dominated by own innovations from the short to the long term accounting for over 99% to about 78% whilst that due to domestic borrowing takes about 12% of the innovations. The contributions of the various innovations to the movements in the indirect tax variable are mainly due to own movements and those that coming from direct taxes. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 19 | P a g e Specifically, own innovations account for over 63% to about 58% from the short to the long period whereas proportion of innovations from direct taxes range from 34% to 30%. Again movements in external borrowing are dominated by own innovations and predictably followed by innovations due to grants. Impulse Response Functions Finally in our analysis, we attempt to trace the effects of shocks emanating from the other variables in the system on each endogenous variable. (Estimations are found in the appendices) We first consider the aggregated government expenditure model. In respect of total government expenditure we realize that its time path around equilibrium is not very much affected by own shocks and that emanating from the other variables. However shocks coming from own innovations and from other variables cause more instability in the trajectory of direct tax variable around the equilibrium path. The instability as witnessed from the graphs is more pronounced especially in response to own shocks and the shocks which originate from total government expenditure and indirect taxes. The greatest effect of any shock in the system on indirect taxes comes from grants. However, the trajectory of external borrowing is affected much more by shocks from grants and then by own shocks than shocks coming from any other variable in the system. For domestic borrowing, apart from shocks triggered from grants the other shocks appear not to have any significant drift in its time path. Finally movements in the grants are largely unaffected by shocks which are transmitted from other variables. It is only own shocks which appear to drift the trajectory of grants from the equilibrium position. In the disaggregated government expenditure model, the story is different from that which is experienced in the aggregated expenditure model. From the graphs, we observe that shocks from government capital expenditure aside of own shocks are those which have more impact on the movement of government consumption expenditure. The time path of government capital expenditure is affected more in the early periods by shocks from government consumption expenditure and own shocks. The shocks from the other variables do not cause as much trepidation. In respect of direct taxes own shocks are the most prominent among all the shocks which are transmitted from the various variables whilst indirect taxes react to own shocks and that which emanates from direct taxes. It is also observed that the effects of shocks from government capital expenditure are noticeable only in the early period of the time horizon. Conclusions and Policy Implications In this study, our major preoccupation has been to establish the nexus between total government expenditure and disaggregated government expenditures and their corresponding financing modes, particularly focusing on the effects of foreign aid well as the response from domestic borrowing. Its import has been to verify whether the theoretical precepts established in the fiscal response models found in the literature still hold true for the Ghanaian economy using current data available. In our analysis we have generally found that whether government expenditure is aggregated or disaggregated, there exists one or other long run equilibrium relationship between expenditure and other variables in the model. More specifically, in the disaggregated government expenditure model, we have found that there only one co integrating equation exists between government consumption expenditure and other variables – government capital expenditure, direct taxes, indirect taxes, external borrowing, domestic borrowing and grants whereas in the model involving aggregate government expenditure, we mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 20 | P a g e discovered two co integrating equations-one for government expenditure and the other for external borrowing. In the long run, we find that external borrowing and grants lead to increased government consumption expenditure but government capital expenditure negatively impacts on government consumption expenditure. The positive effect of external borrowing and grants on government expenditure may point to aid fungibility though that conclusion may be erroneous or flawed on the grounds that some aid and grants come in the form of budgetary support and are therefore legitimately channeled into those areas of government spending which are important in the government's scheme of things. In respect of the aggregate model, external borrowing, domestic borrowing and grants all in the long run lead to increase in government expenditure which confirms concept of aid illusion but surprisingly the domestic revenue streams –direct and indirect taxes trigger a negative response from government expenditure. The estimated long run equation for external borrowing also shows that increased government expenditure precipitates increased external borrowing. Domestic borrowing also has the same effect but direct taxes, indirect taxes and grants all exert a negative effect on external borrowing. The positive effect of domestic borrowing on external borrowing probably gives the indication that because of the inadequacy of the domestically mobilized revenues, external and domestic borrowings have become an important but constant feature of financing government activities. Thus in the long run, in the disaggregated model we were able to adduce evidence of domestic revenues being used to replace external borrowing as a financing avenue. In the literature there is an opinion which articulates the view that external borrowing leads to a lax attitude towards domestic revenue mobilization, usually characterized as the displacement hypothesis. This is partially affirmed by our results in the short run. This is because whilst the effect of external borrowing on direct taxes is positive in both aggregate and disaggregated expenditure models it leads to a decline in indirect taxes in the aggregate model and has an insignificant impact on indirect taxes in the disaggregated model. From the short run results, the external sources of government financing impact positively on the government capital expenditures and this implies these resources are going into areas of the economy which may be reproductive and thus helping to expand economic activities in the long run. In long run it is established that an increased external borrowing and grants lead to more than proportionate growth in aggregate government expenditure which suggests that these external financing channels come with local or counterpart funding components which also exert more pressure on government finances. To ease pressure on government, government would have to enter into external funding agreements which do not require too much of counterpart funding. One other view proffered by some economists in the literature is that governments in developing countries have a preference for grants than loans for financing projects and programmes. In our analysis, it is obvious that the effect of grants undermines direct tax collection and it does appear because grants are normally free, its increased flow into the Ghanaian economy dampens the direct tax collections. Policy makers are encouraged to continue design tax policies and mechanisms which would in spite of increased flow of grant enable the government to rake in the desired revenues. Another significant and illuminating finding is the fact that short run effect of domestic borrowing on both direct and indirect taxes is negative in the Ghanaian economy which signals that domestic borrowing may be inhibiting economic activities and thus may ultimately be having a distortionary mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 21 | P a g e impact on tax collections through its effect on economic activities. The government would therefore do well to scale down on its appetite for borrowing from domestic sources which particularly has a strangulating effect on private sector activities and ultimately impacts negatively on economic activities that generate the revenue needed by government. One of the objectives of the study is to determine whether the borrowing modes have differential effects on the domestic tax channels and our estimated equations suggest that the tax channels do not response in the same way to borrowing. This therefore allows policy makers to design the relevant but right mechanisms to ensure continuous increased tax yields from both direct and indirect sources by creating unique mechanisms which work for each tax channel. Finally we also find that the short-run equation for domestic borrowing in the disaggregated government expenditure model shows that external borrowing is used to substitute domestic borrowing to certain extent and this has a huge implication for the Ghana's debt sustainability which has become a source of worry to international agencies and economic think-tanks within Ghana even against the background of a re-based economy. To conclude we would say that though we have through this study unearthed some important facts relating to the nexus among the fiscal variables and the borrowing modes in Ghana, we would have wished that we were able to segregate aid into the various other forms project, programme or even technical by which they come, which in our view would have enriched the analysis .It is therefore our hope that future studies would tackle this aspect to further add to the existing stock of knowledge in this area. Another area which may be interesting to examine in the future is the effects of these borrowing modes on private investments and economic growth. References Ali, A.A., G.,Malwanda,C., & Sliman, Y. (1999). Official development assistance to Africa: an overview. Journal of African Economies, 8 (4), 504-527. Anaman, E., A., Gadzo, S., G., Gatsi, J., G., & Pobbi, M. (2017). Fiscal aggregates government borrowing and economic growth in Ghana: an ever correction approach. Advances in Management and Applied Economics, 7 (2), 83-104. Bhasin, V.K. (2004). Dynamic interlinks among exchange rate price level and terms of trade ina managed floating exchange system: the case of Ghana. AERC Research Paper, 141. African Economic Consortium, Nairobi, Kenya. Blanchard, O., & Perrotti,R. (1999). An empirical characterization of dynamic effects of changes in government spending and taxes on output. NBER Working Paper, 7269. Bwire,T.,Lloyd, T., &Morrissey, O. (2017). Fiscal reforms and the fiscal effects of aid in Uganda. The Journal of Development Studies, 53 (7), 1019-1036. Deverajan, S., Rajkumar, A., S., & Swaroop, V. (1998). What does aid do to African finance? AERC/ODC Sponsorship on Managing a Smooth Transition from Aid Dependence in Africa, Washington DC. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 22 | P a g e Durdonoo, C. (2000). Fiscal trends: 1970-1995 in E.A. Aryeetey et al (eds.). Economic Reforms in Ghana: the miracle and the mirage. Trenton, USA: Africa Word Press. Enders,W. (1995). Applied econometric time series. New York: Wiley Press. Feyzioglu, T.,Swaroop, V., & Zhu, M. (1998). A panel data analysis of the fungibility of foreign aid. World Bank Economic Review, 65, 429-445. Griffin, K. (1970). Foreign capital domestic savings and economic development. Oxford Bulletin of Economics and Statistics, 55, 99-112. Heller,S., P. (1975). A model of public fiscal behaviour in developing countries: Aid, investment and taxation. American Economic Review, 65 (3), 429-445. Johassen.(1991). Estimation and hypothesis testing of co integration vectors in Gaussian vector autoregressive models. Econometrica Vol59ːpp 1551-1580 Johnston, J.& Di-Nardo, J.(1997). Econometric methods. (4thed.). Singapore: McGraw Hill. Khilji,N.M.,& Zampelli, E., M. (1994). The fungibility of US military and non-military assistance and the impacts on expenditures of major aid recipients. Journal of Development Economics, 43, 345-362. Lloyd, T.,McGillivray, M., Morrissey, O., & Opoku-Afari,M. (2009). The fiscal effects of aid in developing countries: a comparative dynamic analysis. Studies in Development Economics and Policy, 158-179. Lutkepohl, H., & Rimmers, H., E. (1992). Impulse response analysis of co integrated systems. Journal of Economic Dynamics and Control, 16, 53-78. M’Amanja, D.,Lloyd, T., &Morrissey, O. (2005). Fiscal aggregates, aid and growth in Kenya: a vector autoregressive (VAR) analysis. CREDIT Research Paper, 5 (7). Martins, P.M.G. (2010). Fiscal dynamics in Ethiopia: the co integrated VAR model with quarterly data. University of Nottingham CREDIT Research Paper, 10 (5). Mascagni,G., &Timmis, E. (2014). Fiscal effects of aid in Ethiopia: evidence from CVAR application. University of Nottingham CREDIT Research Paper, 14 (6). Mavrotas, G. (2002). Foreign aid and fiscal response: does aid disaggregation matter? Weltwirtschaftliches Archive, 138, 534-559. mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 23 | P a g e McGillivray,M& Morrissey,O. (2000). Aid fungibility in assessing aid: red herring or true concern? Journal of International Development, 12 (3), 413-428. Mosley, P., J., Hudson, J., & Horrell, S. (1987). Aid and public sector and the market in less developed countries. Economic Journal, 97 (9), 616-641. Njeru, J. (2004). The impact of foreign aid on public expenditure: the case of Kenya. AERC Research Paper, 135. African Economic Research Consortium, Nairobi. Osei, R., Morrissey, O., & Lloyd,T. (2003). Modelling the fiscal effects of aid: an impulse response analysis for Ghana. University of Nottingham CREDIT Research Paper, 3 (10). Thomas, R.L. (1993). Introductory econometrics: theory of applications. (2nded.). London: Longman Press. Appendix A 1a Unnormalized Co Integrating Coefficients For Disaggregated Government Expenditure Model LGC LGK LDT LIT LFB LDB LGR C 1.294453 0.343061 1.224174 5.581041 -2.682037 1.252628 -1.990298 50.41932 2.473469 -4.926923 3.323305 0.203476 0.760543 1.084518 1.902956 -9.314356 -2.734475 2.252014 1.876378 -0.711541 -1.777021 -1.152836 -0.274655 58.36492 0.626448 0.708572 -0.757148 -3.873066 -2.603299 4.320096 1.164597 -2.573859 -3.721753 2.645566 2.914102 -2.369617 1.228391 -1.890964 -1.548250 31.02504 1.690610 -1.846113 4.246725 -3.162973 0.280160 -0.497808 0.356670 - 6.716738 2.546556 -2.991196 -0.573203 2.166093 1.917874 -0.844341 -0.672078 - 18.44867 1b Unnormalized Co Integrating Coefficients For Aggregated Government Expenditure Model LGE LDT LIT LFB LDB LGR C 0.331685 2.332985 6.191146 -2.736592 -0.092866 -2.182888 85.96550 -1.536605 5.185054 -5.442559 1.073965 2.058536 3.170925 -61.33498 1.272844 -3.814032 - 0.244547 -3.236886 0.620246 -0.427752 60.13527 -0.924257 2.266703 1.782250 2.664649 -3.404383 0.070304 -34.99253 -0.416011 -2.639002 3.722309 1.874406 0.312871 0.073111 -38.92193 0.081428 -2.878044 0.715399 -0.521188 0.300898 0.115664 8.388109 Appendix B mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 24 | P a g e Short-Run /Error Correction Estimates Short Run Estimates For The Disaggregated Government Expenditure Model Error Correction: DLGC DLGK DLDT DLIT DLFB DLDB DLGR ECT1(-1) - 0.019809 -0.024172 -0.001566 - 0.018902 0.001415 - 0.005638 0.022009 (0.00701) (0.00619) (0.00604) (0.00679) (0.00492) (0.00699) (0.02305) [- 2.82725] [- 3.90412] [- 0.25948] [- 2.78290] [ 0.28745] [- 0.80695] [ 0.95466] DLGC(-1) -0.701589 0.016095 0.155862 0.284686 - 0.044776 0.180555 0.416556 (0.15231) (0.13459) (0.07841) (0.14765) (0.10699) (0.07136) (0.14713) [- 4.60624] [ 0.11959] [ 1.98767] [ 1.92805] [-0.41851] [ 2.53034] [ 2.83116] DLGK(-1) 0.564860 - 0.296333 -0.125927 -0.074017 0.071090 -0.217699 - 0.206533 (0.17444) (0.15415) (0.15030) (0.16911) (0.12254) (0.17396) (0.57399) [ 3.23805] [- 1.92237] [- 0.83782] [- 0.43769] [ 0.58015] [-1.25140] [- 0.35982] DLDT(-1) 0.320703 0.606177 -0.546285 0.319727 - 0.302274 0.032411 0.414471 (0.15278) (0.28556) (0.18010) (0.16165) (0.14998) (0.20845) (0.68779) [ 2.09905] [ 2.12277] [- 3.03324] [ 1.97785] [- 2.01549] [ 0.15549] [ 0.60261] DLIT(-1) 0.186466 -0.566537 -0.001965 - 0.346928 -0.189941 0.065452 - 1.080978 (0.26140) (0.22288) (0.22522) (0.17619) (0.18362) (0.26068) (0.86010) [ 0.71335] [ - 2.54186] [- 0.00873] [- 1.96908] [- 1.03445] [ 0.25109] [- 1.25680] DLFB(-1) 0.098782 0.050804 0.331934 -0.319583 - 0.268403 - 0.132600 1.249801 (0.30483) (0.02321) (0.16903) (0.29551) (0.11911) (0.05443) (1.00302) mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 25 | P a g e [ 0.32405] [ 2.18861] [1.96381] [- 1.08146] [- 2.25347] [- 2.43620] [ 2.24603] DLDB(-1) 0.223057 0.364305 -0.307923 - 0.430906 0.076035 - 0.093922 -1.103910 (0.11371) (0.20497) (0.11830) (0.21693) (0.16293) (0.03904) (0.45124) [1.96165] [ 1.77738] [ - 2.60286] [- 1.98634] [ 0.46667] [- 2.40604] [ - 2.44639] DLGR(-1) 0.203551 -0.100615 - 0.040873 -0.100401 0.120268 0.041354 -0.145643 (0.08956) (0.07914) (0.01616) (0.08682) (0.06099) (0.08931) (0.29467) [ 2.27291] [-1.27141] [-2.52971] [-1.15647] [ 1.97183] [ 0.46305] [- 0.49425] C -0.001451 0.002197 0.001929 0.000621 0.001887 - 0.000925 - 0.063636 (0.04241) (0.03747) (0.03654) (0.04111) (0.02979) (0.04229) (0.13954) [- 0.03421] [ 0.05863] [ 0.05279] [ 0.01511] [ 0.06336] [- 0.02187] [- 0.45605] R-squared 0.551149 0.549121 0.504434 0.583416 0.516661 0.212850 0.526910 Adj. R- squared 0.422906 0.420298 0.362844 0.464393 0.378565 - 0.012050 0.391742 Sum sq. resids 1.859653 1.452124 1.380543 1.747652 0.917605 1.849415 20.13418 S.E. equation 0.257713 0.227731 0.222047 0.249832 0.181029 0.257003 0.847985 F-statistic 4.297695 4.262614 3.562631 4.901676 3.741300 0.946419 3.898175 Log likelihood 2.824038 7.400244 8.335439 3.973200 15.89205 2.926176 -41.24349 Akaike AIC 0.333836 0.086473 0.035922 0.271719 -0.372543 0.328315 2.715864 Schwarz SC 0.725681 0.478318 0.427767 0.663564 0.019302 0.720160 3.107709 Mean dependent - 0.003087 0.001756 0.003245 - 0.001704 0.000154 0.000452 -0.066715 S.D. dependent 0.339245 0.299103 0.278178 0.341370 0.229642 0.255468 1.087286 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 26 | P a g e Appendix C Short Run Estimates For The Aggregated Government Expenditure Model Error Correction: DLGE DLDT DLIT DLFB DLDB DLGR ECT2(-1) -0.814967 -0.559246 -0.693817 0.972253 0.828267 -1.359340 (3.79677) (0.44945) (0.46205) (0.52819) (0.73398) (2.44790) [-3.09437] [-1.24428] [-1.50162] [ 1.84071] [ 1.12847] [-0.55531] ECT3(-1) 0.232244 0.012426 0.353103 -0.425366 -0.111076 0.397426 (0.17271) (0.21543) (0.22146) (0.20943) (0.35180) (1.17329) [ 1.34474] [ 0.05768] [ 1.59442] [-2.03106] [-0.31574] [ 0.33873] DLGE(-1) 0.054739 0.261523 0.573447 0.354344 -0.082371 0.867620 (0.17577) (0.21924) (0.22539) (0.11625) (0.35804) (1.19409) [ 0.31143] [ 1.19284] [ 2.54427] [3..04821] [-0.23006] [ 0.72660] DLGE(-2) -0.090568 0.205657 0.458363 -0.061591 -0.333092 0.248939 (0.16896) (0.21075) (0.21665) (0.24767) (0.34416) (1.14782) [-0.53604] [ 0.97584] [ 2.11564] [-0.24868] [-0.96783] [ 0.21688] DLGE(-3) -0.134650 0.035054 0.129480 0.087230 0.271188 0.231454 (0.06922) (0.15817) (0.16260) (0.18588) (0.13594) (0.86147) [-1.97187] [ 0.22162] [ 0.79629] [ 0.46927] [ 1.99489] [ 0.26867] DLDT(-1) 0.172514 0.050142 0.517520 0.256857 0.033832 -2.202289 (0.23221) (0.28965) (0.18901) (0.34039) (0.47301) (1.10333) [ 0.74293] [ 0.17311] [ 2.73802] [ 0.75459] [ 0.07153] [-1.99603] DLDT(-2) 1.148476 0.025518 -0.138434 -0.536571 -0.845334 2.062187 (0.25936) (0.32352) (0.33258) (0.38020) (0.52832) (1.76201) [ 4.42808] [ 0.07888] [-0.41624] [-1.41130] [-1.60004] [ 1.17036] DLDT(-3) 0.349772 0.069256 0.305151 -0.916470 -1.702434 3.984505 (0.35723) (0.03213) (0.45809) (0.46590) (0.72769) (1.50826) [ 0.97911] [ 2.15542] [ 0.66614] [-1.96709] [-2.33951] [ 2.64179] DLIT(-1) 0.553068 0.253030 -0.338736 -1.382197 -1.282164 2.899246 (0.37157) (0.46348) (0.47647) (0.54468) (0.64301) (2.52431) [ 1.48846] [ 0.54593] [-0.71093] [-2.53762] [-1.99399] [ 1.14853] mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 27 | P a g e DLIT(-2) -0.395099 0.330036 0.236805 -0.224872 -0.091448 -0.446796 (0.17676) (0.16527) (0.12177) (0.25911) (0.36005) (1.20082) [-2.23527] [ 1.99690] [ 1.94477] [-0.86787] [-0.25398] [-0.37208] DLIT(-3) -0.234890 -0.094442 -0.243058 0.104862 0.414603 -1.457794 (0.18283) (0.22806) (0.23445) (0.26802) (0.37243) (1.24211) [-1.28471] [-0.41411] [-1.03671] [ 0.39125] [ 1.11323] [-1.17364] DLFB(-1) 0.115739 -0.055537 -0.337443 0.523889 0.179581 0.007247 (0.36633) (0.45694) (0.46975) (0.26518) (0.74621) (2.48870) [ 0.31594] [-0.12154] [-0.71835] [ 1.97559] [ 0.24066] [ 0.00291] mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 28 | P a g e DLFB(-2) 0.224531 0.424402 -0.240696 -0.437316 -0.682534 (0.09086) (0.16517) (0.08399) (0.31780) (0.44162) [ 2.47116] [ 2.56939] [-2.86581] [-1.37606] [-1.54553] DLFB(-3) 0.150271 0.119247 -0.043971 0.000501 -0.094436 (0.20228) (0.25232) (0.25939) (0.29653) (0.41205) [ 0.74287] [ 0.47260] [-0.16952] [ 0.00169] [-0.22919] DLDB(-1) 0.337644 -0.223752 0.276325 0.308478 0.447370 (0.13910) (0.29508) (0.30334) (0.34677) (0.48187) [2.42731] [-0.75829] [ 0.91093] [ 0.88957] [ 0.92840] DLDB(-2) 0.259235 -0.424442 -0.196132 0.022124 -0.162952 (0.18529) (0.21368) (0.23760) (0.27162) (0.37744) [1.39905] [-1.98639] [-0.82545] [ 0.08145] [-0.43173] DLDB(-3) 0.087809 -0.126465 0.133979 0.311050 0.399951 (0.16787) (0.20940) (0.21527) (0.24609) (0.18434) [ 0.52306] [-0.60394] [ 0.62238] [ 1.26399] [ 2.16959] DLGR(-1) 0.969935 -0.641608 -0.430083 1.049242 1.055592 (0.36425) (0.45435) (0.46708) (0.53395) (0.74197) [2.66285] [-1.41215] [-0.92079] [ 1.96507] [ 1.42269] DLGR(-2) 0.505786 -0.464436 -0.363172 0.510144 0.410434 (0.23730) (0.23587) (0.30429) (0.34786) (0.48338) [2.13141] [-1.96904] [-1.19349] [ 1.46653] [ 0.84909] DLGR(-3) 0.070551 -0.231593 -0.088270 0.314585 0.177093 (0.12391) (0.15457) (0.15890) (0.11515) (0.25241) [0.56935] [-1.49833] [-0.55551] [ 2.73185] [ 0.70159] C 0.004849 0.012664 0.020949 -0.005055 -0.014179 (0.02025) (0.02526) (0.02597) (0.02969) (0.04125) [ 0.23941] [ 0.50132] [ 0.80665] [-0.17026] [-0.34369] R-squared 0.927544 0.778822 0.895330 0.783659 0.654236 Adj. R-squared 0.824036 0.462853 0.745801 0.474601 0.160288 1.368240 (1.47284) [ 0.92898] 2.359789 (1.18157) [ 1.99717] 0.735352 (1.60710) [ 0.45756] 0.261685 (1.25881) [ 0.20788] -0.663780 (1.14047) [-0.58202] -2.475217 (2.47456) [-1.00027] -1.610878 (1.61213) [-0.99922] -0.333660 (0.84183) [-0.39635] -0.076360 (0.13759) [-0.55499] 0.790353 0.490856 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 29 | P a g e Sum sq. resids 0.189327 0.294576 0.311316 0.406836 0.785588 S.E. equation 0.116290 0.145056 0.149120 0.170469 0.236883 F-statistic 8.961041 2.464869 5.987662 2.535638 1.324503 Log likelihood 41.68061 33.94453 32.97730 28.29429 16.77889 Akaike AIC -1.181749 -0.739688 -0.684417 -0.416817 0.241207 Schwarz SC -0.248541 0.193521 0.248792 0.516392 1.174415 Mean dependent 0.002435 0.012086 0.015653 0.003679 - 0.002856 S.D. dependent 0.277223 0.197919 0.295767 0.235180 0.258505 8.738089 0.790031 2.638939 -25.37886 2.650220 3.583429 -0.076870 1.107195 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 30 | P a g e Variance Decomposition of DLGE: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR 1 0.111813 100.0000 0.000000 0.000000 0.000000 0.000000 0.000000 2 0.265269 23.43584 20.37757 3.577401 0.422971 0.413212 51.77301 3 0.385462 15.65435 12.85675 2.279425 1.161694 0.582181 67.46560 4 0.444459 15.27655 11.13926 1.777144 0.909225 4.016351 66.88147 5 0.480041 18.49762 9.680806 2.009450 0.842194 3.446230 65.52370 6 0.509495 18.37302 9.024340 2.928824 2.621776 3.661696 63.39034 7 0.530525 18.10874 8.592325 2.945605 2.653703 3.634360 64.06527 8 0.592717 16.41770 9.398415 2.439467 2.725268 3.237613 65.78154 9 0.645311 16.74329 8.484315 2.063404 2.578772 3.102590 67.02763 10 0.707590 17.26560 7.260408 2.026676 2.782769 2.649450 68.01510 Variance Decomposition of DLDT: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR 1 0.135609 14.67399 85.32601 0.000000 0.000000 0.000000 0.000000 2 0.168781 18.62678 57.94360 15.88752 2.083583 5.320575 0.137940 3 0.171894 18.51302 56.19375 15.58305 2.923035 6.649722 0.137425 4 0.187967 15.56514 56.71454 16.41705 2.446050 6.768987 2.088241 5 0.196661 15.84631 52.83267 19.01200 2.663066 6.243143 3.402820 6 0.201969 15.12285 50.18036 18.64150 3.122657 5.935793 6.996830 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 31 | P a g e 7 0.202326 15.12402 50.07573 18.69772 3.124804 5.957755 7.019973 8 0.206023 14.68438 49.98696 18.16153 3.086644 7.243165 6.837325 9 0.207504 15.14699 49.39309 18.35866 3.175843 7.150039 6.775385 10 0.210232 15.37626 48.35545 17.94419 3.096044 6.967732 8.260324 Variance Decomposition of DLIT: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR 1 0.172417 1.586740 48.92190 49.49136 0.000000 0.000000 0.000000 2 0.222251 6.532833 39.76821 36.29187 2.542720 4.247651 10.61671 3 0.240533 5.666080 42.98798 32.38567 2.495584 3.642069 12.82262 4 0.282419 7.145570 32.56612 24.49860 1.925561 2.648976 31.21517 5 0.305613 7.189801 30.23615 20.92503 1.708072 3.511665 36.42929 6 0.324604 9.538451 27.59595 18.58652 1.582996 3.176390 39.51969 7 0.340375 11.50168 25.84900 19.30786 1.613194 2.956258 38.77200 8 0.349941 12.00197 24.45653 18.30123 1.805046 3.433497 40.00173 9 0.372219 12.16112 23.06242 16.47424 2.106031 3.147840 43.04836 10 0.392478 12.20646 21.59352 14.82943 2.053484 3.118234 46.19888 Variance Decomposition of DLFB: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 32 | P a g e 1 0.158397 0.044068 1.797966 0.832714 97.32525 0.000000 0.000000 2 0.177824 0.283293 10.59105 4.897140 78.39354 0.021085 5.813892 3 0.231639 1.864778 7.092211 2.953018 60.72698 1.889828 25.47319 4 0.245491 1.801467 6.655475 7.144054 56.78387 3.317991 24.29714 5 0.256219 3.597538 6.326243 7.361394 52.13185 3.170792 27.41218 6 0.265092 5.934772 5.950920 7.822540 49.15187 3.312355 27.82755 7 0.276468 6.407326 6.361810 7.229981 45.20437 3.128504 31.66801 8 0.292901 7.634311 6.580809 6.595101 40.28310 3.123580 35.78309 9 0.302819 8.628087 6.568402 6.291867 37.78864 3.198804 37.52420 10 0.311030 9.505549 6.228335 6.097690 36.03080 3.074052 39.06357 Variance Decomposition of DLDB: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR 1 0.206871 2.265834 0.810045 2.773664 22.89687 71.25359 0.000000 2 0.223675 2.783841 6.202074 4.149667 19.80236 64.21923 2.842822 3 0.288792 3.716311 3.763684 2.544012 22.32057 38.78066 28.87476 4 0.338815 12.26334 2.735667 1.852837 16.64807 31.38268 35.11741 5 0.401912 19.12568 2.211283 6.381911 11.92865 23.45652 36.89596 6 0.441166 19.68044 2.240487 6.544079 11.51591 19.74382 40.27526 7 0.461121 19.85563 3.003299 5.993447 10.54970 18.09470 42.50323 8 0.502054 17.95271 5.140998 5.121903 9.264051 15.39093 47.12941 9 0.537271 16.97087 5.065356 4.512318 9.047540 13.47473 50.92919 10 0.574863 16.64304 4.665101 3.945109 8.396938 11.83882 54.51099 Variance Decomposition of DLGR: Period S.E. DLGE DLDT DLIT DLFB DLDB DLGR 1 0.755211 0.207621 2.294906 4.433442 39.47587 2.841255 50.74691 2 0.892897 5.268719 2.287596 3.424055 40.66342 10.57656 37.77965 3 1.118448 6.441865 2.333939 4.617557 28.37545 6.795336 51.43585 4 1.162192 11.41789 2.963043 4.304557 26.28069 6.318548 48.71527 5 1.250068 12.21659 4.194314 4.814527 22.72084 5.506416 50.54731 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 33 | P a g e 6 1.335245 12.91167 4.548363 4.353771 20.51086 4.894152 52.78119 7 1.416154 13.08263 5.335398 3.877774 18.46247 4.949913 54.29181 8 1.525871 13.21428 4.937466 3.401963 16.37939 4.283290 57.78361 9 1.618501 13.31405 4.747701 3.062418 15.41920 3.826015 59.63062 10 1.693009 14.17167 4.431628 2.800191 14.16598 3.500881 60.92966 Variance Decomposition of Disaggregated Model Variance Decomposition of DLGC: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.21708 5 100.000 0 0.00000 0 0.00000 0 0.00000 0 0.00000 0 0.00000 0 0.00000 0 2 0.26964 9 65.0425 8 23.12333 1.721511 5.952523 1.337863 2.821443 0.00074 6 3 0.27782 6 61.27276 21.81767 1.692895 7.846924 1.277528 3.468516 2.623707 4 0.27882 7 60.9569 0 21.66242 1.705474 7.793662 1.791860 3.467854 2.621829 5 0.279151 60.8222 2 21.63173 1.708604 7.792124 1.787999 3.499983 2.757338 6 0.27924 8 60.78512 21.62981 1.708247 7.786927 1.810778 3.523712 2.755411 7 0.27926 2 60.7797 8 21.62889 1.708456 7.786194 1.810644 3.527721 2.758316 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 34 | P a g e 8 0.27926 6 60.77815 21.62924 1.708446 7.786028 1.811031 3.528838 2.758263 9 0.27926 6 60.7780 3 21.62921 1.708450 7.786039 1.811035 3.528958 2.758272 10 0.27926 6 60.7780 0 21.62923 1.708450 7.786037 1.811035 3.528974 2.758273 Variance Decompositio n of DLGK: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.22126 0 28.6533 7 71.34663 0.00000 0 0.00000 0 0.00000 0 0.00000 0 0.00000 0 2 0.25624 1 22.7309 3 54.92869 2.311531 17.06961 0.387709 0.28668 8 2.284841 3 0.25960 9 22.6632 9 53.98672 2.253508 16.75220 1.325450 0.502436 2.516391 4 0.26065 2 22.49917 53.67423 2.237255 16.72579 1.333753 0.538337 2.991464 5 0.26092 9 22.4750 5 53.57948 2.234198 16.69179 1.418543 0.615461 2.985473 6 0.26098 4 22.4669 2 53.56351 2.234675 16.68504 1.418000 0.630732 3.001125 7 0.26100 0 22.46491 53.56006 2.234580 16.68338 1.420510 0.635649 3.000918 8 0.26100 1 22.4647 0 53.55959 2.234606 16.68323 1.420555 0.636214 3.001109 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 35 | P a g e 9 0.26100 2 22.4646 3 53.55954 2.234606 16.68319 1.420579 0.636324 3.001121 10 0.26100 2 22.4646 3 53.55953 2.234607 16.68320 1.420583 0.63633 0 3.001121 Variance Decompositio n of DLDT: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.177084 0.109107 0.077847 99.81305 0.000000 0.000000 0.000000 0.000000 2 0.200855 1.745636 2.080662 80.25215 0.882415 0.001371 12.52711 2.510653 3 0.202226 1.821304 2.507557 79.17233 1.231177 0.137470 12.52452 2.605643 4 0.202403 1.852392 2.556090 79.03863 1.231784 0.193493 12.52652 2.601091 5 0.202479 1.851029 2.572860 78.98032 1.242394 0.193999 12.52149 2.637904 6 0.202498 1.852722 2.572986 78.96633 1.243655 0.201673 12.52373 2.638906 7 0.202501 1.852663 2.573234 78.96395 1.243682 0.202162 12.52383 2.640481 8 0.202502 1.852679 2.573349 78.96325 1.243719 0.202460 12.52393 2.640620 9 0.202502 1.852677 2.573347 78.96319 1.243719 0.202496 12.52392 2.640650 10 0.202502 1.852677 2.573353 78.96317 1.243719 0.202502 12.52392 2.640656 Variance Decomposition of DLIT: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.21264 6 1.939034 0.55468 2 34.35595 63.15033 0.00000 0 0.00000 0 0.00000 0 2 0.23307 0 5.509995 0.50597 8 30.8045 7 59.4287 0 0.027701 3.722568 0.00049 2 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 36 | P a g e 3 0.23484 3 5.627085 1.376441 30.35335 58.5369 3 0.222132 3.827006 0.057054 4 0.23528 6 5.608014 1.415256 30.2403 0 58.3827 9 0.250664 3.825924 0.277053 5 0.23538 7 5.60929 0 1.414108 30.21875 58.3383 8 0.295249 3.830620 0.293600 6 0.23540 8 5.608271 1.414022 30.21479 58.3287 0 0.299418 3.830326 0.304476 7 0.235414 5.608157 1.414510 30.21367 58.3260 5 0.301472 3.830780 0.305368 8 0.235415 5.608128 1.414509 30.21357 58.3257 2 0.301660 3.830780 0.305632 9 0.235415 5.608121 1.414552 30.21354 58.3256 4 0.301701 3.830795 0.305656 10 0.235415 5.608120 1.414554 30.21353 58.3256 3 0.301705 3.830795 0.305658 Variance Decompositio n of DLFB: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.156790 0.00078 0 7.993319 4.30490 4 6.65E- 05 87.70093 0.00000 0 0.00000 0 2 0.178720 0.612045 9.62814 8 4.875545 2.49460 7 67.50957 0.08868 2 14.79141 3 0.184871 0.773861 10.6418 0 4.59492 6 2.99994 6 65.36298 1.283818 14.34266 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 37 | P a g e 4 0.185743 0.95058 4 10.5569 4 4.56668 4 2.98130 2 64.84869 1.658247 14.43755 5 0.18608 0 0.955757 10.6652 2 4.550185 2.97332 0 64.63541 1.794920 14.42519 6 0.186132 0.961358 10.6650 8 4.547643 2.97704 5 64.60594 1.825808 14.41713 7 0.186140 0.961336 10.6676 4 4.547317 2.97678 9 64.6000 8 1.829933 14.41690 8 0.186142 0.961392 10.6676 9 4.54726 0 2.97703 0 64.59911 1.830600 14.41691 9 0.186142 0.961391 10.6676 9 4.547263 2.97703 2 64.59909 1.830626 14.41691 10 0.186142 0.961391 10.6676 9 4.547263 2.97703 7 64.59906 1.830627 14.41692 Variance Decompositio n of DLDB: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.215948 4.175000 4.859923 0.489323 0.605885 22.17905 67.69082 0.000000 2 0.234199 3.952763 7.392635 1.278903 0.638402 18.86131 65.22346 2.652535 3 0.237645 4.077868 7.873997 1.298188 0.911934 18.73193 64.46117 2.644914 4 0.237988 4.068213 7.914957 1.301817 0.927342 18.70947 64.41447 2.663731 5 0.238062 4.065940 7.924473 1.302481 0.934178 18.70143 64.39396 2.677537 6 0.238067 4.065806 7.924171 1.302719 0.934944 18.70362 64.39132 2.677422 7 0.238068 4.065833 7.924185 1.302764 0.934989 18.70346 64.39078 2.677982 8 0.238069 4.065830 7.924253 1.302763 0.935003 18.70350 64.39066 2.677997 9 0.238069 4.065836 7.924250 1.302763 0.935004 18.70349 64.39065 2.678003 10 0.238069 4.065835 7.924255 1.302763 0.935004 18.70349 64.39065 2.678004 mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 38 | P a g e Variance Decomposition of DLGR: Period S.E. DLGC DLGK DLDT DLIT DLFB DLDB DLGR 1 0.755153 0.002803 11.06231 1.502395 2.097732 16.00708 9.021366 60.30632 2 0.822087 0.060881 12.26010 1.790907 3.884525 21.34494 9.482869 51.17578 3 0.829529 0.293374 12.52555 2.002059 3.815918 20.96703 9.393659 51.00241 4 0.832194 0.295626 12.77577 1.997428 3.841262 20.91620 9.436867 50.73685 5 0.832515 0.311644 12.76592 1.996002 3.850621 20.91331 9.464671 50.69783 6 0.832614 0.311572 12.77297 1.995730 3.850479 20.90860 9.469079 50.69158 7 0.832635 0.311966 12.77253 1.995715 3.851079 20.90871 9.470205 50.68979 8 0.832637 0.311967 12.77253 1.995745 3.851055 20.90879 9.470231 50.68968 9 0.832638 0.311969 12.77252 1.995752 3.851068 20.90879 9.470237 50.68966 10 0.832638 0.311969 12.77252 1.995754 3.851067 20.90880 9.470235 50.68965 Cholesky Ordering DLGC DLGK DLDT DLIT DLFB DLDB DLGR mailto:contact@americaserial.com mailto:contact@americaserial.com American Research Journal of Economics, Finance and Management Volume 11 Issue 2, April-June 2023 ISSN: 2836-9416 Impact Factor: 5.57 Journal Homepage: https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com Official Journal of America Serial Publication American Research Journal of Economics, Finance and Management https://americaserial.com/Journals/index.php/ARJEFM, Email: contact@americaserial.com 39 | P a g e IMPULSE RESPONSE FUNCTIONS FOR AGGREGATED GOVERNMENT EXPENDITURE MODEL Response to Cholesky One S.D. Innovations ± 2 S.E. MODEL FUNCTIONS Response to Cholesky One S.D. Innovations ± 2 S.E. mailto:contact@americaserial.com mailto:contact@americaserial.com